@INCOLLECTION{CREPALDI_2022_INCOLLECTION_CFMNPT_471258, AUTHOR = {Crepaldi, D. and Ferro, M. and Marzi, C. and Nadalini, A. and Pirrelli, V. and Taxitari, L.}, TITLE = {Finger movements and eye movements during adults' silent and oral reading}, YEAR = {2022}, ABSTRACT = {Using a common tablet and a web application, we can record the finger movements of a reader that is concurrently reading and finger-pointing a text displayed on the tablet touchscreen. In a preliminary analysis of "finger-tracking" data of early-graders we showed that finger movements can replicate established reading effects observed in more controlled settings. Here, we analyse and discuss reading evidence collected by (i) tracking the finger movements of adults reading a short essay displayed on a tablet touchscreen, and (ii) tracking the eye movements of adults reading a comparable text displayed on the screen of a computer. Texts in the two conditions were controlled for linguistic complexity and page layout. In addition, we tested adults' comprehension in both silent and oral reading, by asking them multiple-choice questions after reading each text. We show and discuss the reading evidence that the two (optical and tactile) protocols provide, and to what extent they show comparable effects. We conclude with some remarks on the importance of ecology and portability of protocols for large-scale collection of naturalistic reading data.}, KEYWORDS = {Reading, finger-tracking, digital technology}, PAGES = {443-471}, URL = {https://link.springer.com/book/9783030998905}, VOLUME = {23}, PUBLISHER = {Springer (Dordrecht, NLD)}, ISBN = {978-3-030-99890-5}, BOOKTITLE = {Developing language and literacy-Studies in Honor of Dorit Diskin Ravid}, EDITOR = {Levie, R. and Bar On, A. and Ashkenazi, O. and Dattner, E. and Brandes, G.}, } @INPROCEEDINGS{MARZI_2022_INPROCEEDINGS_MNFMMVPTP_471602, AUTHOR = {Marzi, C. and Narzisi, A. and Ferro, M. and Masi, G. and Milone, A. and Viglione, V. and Pelagatti, S. and Tomassini, I. and Pirrelli, V.}, TITLE = {Patterns of finger-tracking in Italian early readers with Autism Spectrum Disorder}, YEAR = {2022}, ABSTRACT = {Background: Of late, the synergistic interaction of eye and hand movements in the exploration of a visual scene displayed on a computer touchscreen was shown to provide a congruent signature of the "attention maps" of subjects with autism spectrum disorders (ASD). A familiar context where this visual and tactile interaction is exploited is when children use the finger of their dominant hand to point the letters of written words as they are reading, particularly at early stages of their literacy development. In the present work, a dedicated app running on a common tablet is used to capture and analyse the finger-tracking behaviour of children with ASD while they are reading few episodes of a connected text on the tablet touchscreen. The reader's voice is also recorded through the tablet built-in microphone. The sliding movements of the finger across the tablet touchscreen are discretized into a series of densely distributed "touch events", which are then mapped onto the text lines in much the same way eye fixations are projected onto a sequence of words using an eye-tracker. Reading texts are linguistically annotated, to control for levels of reading difficulty, and finger-tracking times are associated with linguistic glosses. Objectives: Investigate patterns of finger-tracking as a potential non biological marker for identification of children with ASD . Methods: A preliminary analysis is offered of evidence of the finger-tracking behaviour of 20 Italian children with high functioning ASD, aged 7-11 years, while they are engaged in reading. A grade-matched control group of children with typical development was included. Patterns of finger-tracking are assessed in connection with three complementary aspects of reading behaviour: (1) word recognition, (2) pace of reading of multi-word intonation units, and (3) text comprehension, controlled by asking children a few multiple-choice questions on text content after each reading session. Results: Considerable variation in levels of reading ability was observed in the ASD sample, with a few children showing clear evidence of impaired reading comprehension. However, fluent readers with ASD exhibit the same correlation between accurate decoding (assessed by measuring per-word reading speed) and high levels of reading comprehension found in controls. Likewise, decoding rates were found to significantly increase with increasing grade levels, following the typical developmental pattern observed in controls. On a less local level of linguistic analysis, the reading pace of ASD readers fails to be modulated according to major syntactic structures, punctuation marks and direct speech turns, an effect concomitant with a flat prosodic intonation of oral reading. Conclusions: Preliminary findings confirm the heterogeneous nature of reading skills in children with ASD, showing that the use of a tablet screen as a tactile interface for visual perception analysis can offer a robust experimental protocol for large-scale, multimodal collection of naturalistic data for extensive assessment of readers with ASD.}, KEYWORDS = {reading, autism, finger-tracking, developing readers, prediction-driven processing}, PAGES = {192-192}, URL = {https://cdn.ymaws.com/www.autism-insar.org/resource/resmgr/files/insar_2022/2022_Abstract_Book.pdf}, VOLUME = {2022}, CONFERENCE_NAME = {INSAR}, CONFERENCE_PLACE = {Austin, Texas}, CONFERENCE_DATE = {11-14/05/2022}, BOOKTITLE = {2022 annual meeting abstract book}, } @INCOLLECTION{CAPPA_2021_INCOLLECTION_CFG_461297, AUTHOR = {Cappa, C. and Ferro, M. and Giulivi, S.}, TITLE = {Valutare l'efficienza di lettura in classe, fra "ecologia" e tecnologie}, YEAR = {2021}, ABSTRACT = {La sperimentazione AEREST ha consentito la creazione di un protocollo in grado di offrire una valutazione accurata e dettagliata delle abilità di lettura e comprensione del testo. Lo strumento si è rivelato semplice da utilizzare per gli insegnanti, ed è stato accolto con curiosità e interesse dagli allievi, certamente attratti dal supporto utilizzato per la somministrazione (il tablet), ma anche dai testi, che sono stati scelti e adattati con particolare cura. L'analisi dei dati ha consentito di identificare una considerevole varietà di profili di lettori, per i quali sarà possibile progettare percorsi di potenziamento mirati. Come già accennato, si è potuta constatare l'efficacia dello strumento nell'identificazione di allievi le cui difficoltà (pur evidenti agli occhi degli insegnanti) non vengono rilevate dai test comunemente utilizzati per la valutazione, ma la cui lettura non può essere considerata 'efficiente'. Costituiscono esempi in questo senso gli allievi che decodificano in modo accurato e veloce, con buone prestazioni nella comprensione all'ascolto, ma che manifestano difficoltà nella comprensione di un testo in lettura silente, poiché in questa attività devono integrare la decodifica con l'accesso al significato. Un ulteriore esempio è costituito dagli allievi che ottengono buoni risultati in tutti i test, impiegando però un tempo eccessivamente lungo per svolgerli. Nella prospettiva qui adottata, anche per questi allievi è necessario individuare strategie di supporto volte a evitare che le attività scolastiche, in particolare i compiti a casa, occupino una parte troppo ampia del tempo dell'allievo, togliendo spazio al gioco, allo svago, agli interessi personali e alla socializzazione. Questi ultimi sono aspetti che, come sottolinea la Carta internazionale dei diritti dei bambini (1959), rivestono un'importanza cruciale per il processo di crescita e il benessere generale di ciascuno. Oltre alle difficoltà, il protocollo AEREST consente di mettere in evidenza le prestazioni eccellenti, grazie alla struttura dei test e alle caratteristiche dei testi e delle domande che li accompagnano. Capire a fondo come "funzionano" gli allievi è indispensabile per poterli sostenere al meglio negli apprendimenti, indipendentemente dalla presenza o meno di un'"etichetta" diagnostica. Gli insegnanti hanno in questo senso una grande responsabilità, e uno strumento come AEREST, grazie anche all'implementazione su piattaforma tecnologica, può aiutarli in quella che forse è la loro principale sfida quotidiana: fare in modo che le difficoltà scolastiche non siano vissute come barriere all'apprendimento, al successo scolastico, alle opportunità professionali, alla realizzazione personale, ma come soglie da superare e da trasformare in trampolini di lancio.}, KEYWORDS = {efficienza di lettura, decodifca, comprensione, scuola primaria}, PAGES = {49-69}, URL = {https://buponline.com/prodotto/disturbi-specifici-dellapprendimento-e-insegnamento-linguistico/}, VOLUME = {3}, PUBLISHER = {Bononia University Press (Bologna, ITA)}, ISBN = {978-88-6923-829-1}, BOOKTITLE = {Didattica dell'italiano}, EDITOR = {Garulli, V. and Pasetti, L. and Viale, M.}, } @INPROCEEDINGS{BRUNO_2021_INPROCEEDINGS_BGCMF_461393, AUTHOR = {Bruno, E. and Giulivi, S. and Cappa, C. and Marini, M. and Ferro, M.}, TITLE = {Evaluating the accuracy of decoding in children who read aloud}, YEAR = {2021}, ABSTRACT = {Digital tools based on automatic speech recognition (ASR) could be a useful support for teachers in assessing the reading skills of the students. We focus on the evaluation of the decoding accuracy of children with grade level ranging from the 3rd to the 6th performing a reading aloud task on a narrative text displayed on an ordinary tablet using the ReadLet platform. On the basis of previously collected data, we built a gold dataset with sentences characterised by the audio data, the original text to be read, and the text actually spoken by the child. By using the open-source Kaldi toolkit an ASR system based on the GMM-HMM model was trained on the training portion of the gold dataset. The accuracy of the ASR system was calculated as the ability to correctly decode the test audio data with respect to the annotated text, and the decoding accuracy of the children was estimated by measuring the gap between the results obtained with the annotated text and the original text. A consistent trend with increasing grade level was found in terms of word correctness, substitutions and insertions, while the trained model appears to be significantly able to evaluate the children decoding accuracy.}, KEYWORDS = {speech recognition, decoding accuracy, reading aloud, voice parameters, Kaldi, GMM-HMM acoustic model}, PAGES = {145-148}, URL = {https://publications.cnr.it/doc/461393}, DOI = {10.36253/978-88-5518-449-6}, PUBLISHER = {Firenze University Press (Firenze, ITA)}, ISBN = {978-88-5518-449-6}, CONFERENCE_NAME = {12th International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA'21)}, CONFERENCE_PLACE = {Firenze (Italy)}, CONFERENCE_DATE = {14-16/12/2021}, BOOKTITLE = {Proceedings of the 12th International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA'21)}, EDITOR = {Manfredi, C.}, } @INPROCEEDINGS{TAXITARI_2021_INPROCEEDINGS_TCFMNP_441870, AUTHOR = {Taxitari, L. and Cappa, C. and Ferro, M. and Marzi, C. and Nadalini, A. and Pirrelli, V.}, TITLE = {Using mobile technology for reading assessment}, YEAR = {2021}, ABSTRACT = {The enormous potential of Information and Communication Technologies (ICT) for addressing critical educational issues is generally acknowledged, but its use in the assessment of the complex skills of reading and understanding a text has been very limited to date. The paper contrasts traditional reading assessment protocols with ReadLet, an ICT platform with a tablet front-end, designed to support online monitoring of silent and oral reading abilities in early graders. ReadLet makes use of cloud computing and mobile technology for large-scale data collection and allows the time alignment of the child's reading behaviour with texts tagged using Natural Language Processing (NLP) tools. Initial findings replicate established benchmarks from the psycholinguistic literature on reading in both typically and atypically developing children, making the application a new ground-breaking approach in the evaluation of reading skills. Index Terms--reading assessment, reading research, mobile technology, NLP, cloud computing, special education needs.}, KEYWORDS = {reading assessment, reading research, mobile technology, NLP, cloud computing, special education needs}, PAGES = {1-6}, URL = {http://www.ieee.ma/cist20/component/content/?id=26\&Itemid=185}, ISBN = {9781728166469}, CONFERENCE_NAME = {6th IEEE Congress on Information Science \& Technology (IEEE CIST'20)}, CONFERENCE_PLACE = {online}, CONFERENCE_DATE = {05/06/2021}, } @INPROCEEDINGS{MARZI_2021_INPROCEEDINGS_MTFNP_445743, AUTHOR = {Marzi, C. and Taxitari, L. and Ferro, M. and Nadalini, A. and Pirrelli, V.}, TITLE = {Valutare la lettura "in tempo reale": un esempio di integrazione tra linguistica computazionale e linguistica applicata}, YEAR = {2021}, ABSTRACT = {In anni recenti, linguistica computazionale e linguistica applicata hanno ampliato i loro rispettivi ambiti d'indagine, utilizzando l'ontologia formale della linguistica teorica e i modelli cognitivi della psicolinguistica per studiare le difficoltà che i parlanti incontrano nello svolgimento di "compiti" linguistici specifici. Nell'ambito della lettura, le tecnologie per il Trattamento Automatico del Linguaggio (TAL) si sono dimostrate capaci di classificare il livello di leggibilità di un testo, basandosi sulla distribuzione di alcuni parametri linguistici in testi pre-classificati per età dei lettori destinatari, o per grado di scolarità, o per livello di sviluppo cognitivo. Ad esempio, parole o frasi più lunghe, o parole più rare tendono a distribuirsi in testi di più difficile comprensione, o destinati a lettori più maturi. E' possibile così assegnare a un testo, o a ogni singola frase, un punteggio di leggibilità in funzione (inversa) della complessità lessicale, morfologica, sintattica o pragmatica dell'unità testuale analizzata. In Linguistica Applicata (LA) la valutazione della difficoltà di lettura ha seguito un approccio funzionale. Nel modello semplice di lettura, ad esempio, la capacità di leggere un testo è analizzata come il prodotto dell'interazione tra decodifica e comprensione. Attraverso l'osservazione di un campione di bambini impegnati nella lettura, è possibile valutare la loro fluenza in decodifica, gli errori di decodifica e comprensione, e l'efficacia di percorsi educativi personalizzati. La piattaforma ReadLet è stata sviluppata con l'obiettivo di integrare l'approccio classificatorio del TAL con quello funzionale della LA. Il bambino legge un breve testo visualizzato sullo schermo di un tablet, ad alta voce o in modalità silente. In entrambi i casi, al bambino viene chiesto di "tenere il segno" con il dito sullo schermo nel corso della lettura. La traccia tattile è registrata e allineata con il testo visualizzato sullo schermo mediante un algoritmo di convoluzione. Al contempo, il testo è annotato automaticamente per tratti linguistici. Alla fine della sessione di lettura silente, il bambino risponde ad alcune semplici domande sul contenuto del testo. I dati raccolti consentono di valutare le difficoltà (rallentamenti o errori) che il bambino incontra nella lettura, e di mettere in relazione "in tempo reale" queste difficoltà con aspetti linguistici specifici del testo. Un'analisi preliminare dei dati raccolti da ReadLet su oltre 400 allievi di alcune scuole elementari toscane e della Svizzera italiana, ha evidenziato il differente "passo" di lettura tra lettori con sviluppo tipico e atipico, e il peso che variabili come lunghezza, frequenza e lessicalità hanno su profili di lettura individuali e aggregati. La possibilità di "controllare" automaticamente la distribuzione di queste variabili nel testo e di correlarle con le difficoltà del singolo bambino consente, infine, di somministrare testi con livelli di difficoltà gradualmente crescenti, rendendo possibili percorsi personalizzati di potenziamento.}, KEYWORDS = {reading assessment, reading strategies, NLP, ICT mobile technologies}, PAGES = {5-5}, URL = {https://publications.cnr.it/doc/445743}, VOLUME = {2021}, CONFERENCE_NAME = {XXI Congresso Internazionale di AItLA}, CONFERENCE_PLACE = {Bergamo (I)}, CONFERENCE_DATE = {11-12/02/2021}, BOOKTITLE = {FARE LINGUISTICA APPLICATA CON LE DIGITAL HUMANITIES}, } @INCOLLECTION{PIRRELLI_2020_INCOLLECTION_PMFCBM_421741, AUTHOR = {Pirrelli, V. and Marzi, C. and Ferro, M. and Cardillo, F. A. and Baayen, H. R. and Milin, P.}, TITLE = {Psycho-computational modelling of the mental lexicon}, YEAR = {2020}, ABSTRACT = {Over the last decades, a growing body of evidence on the mechanisms governing lexical storage, access, acquisition and processing has questioned traditional models of language architecture and word usage based on the hypothesis of a direct correspondence between modular components of grammar competence (lexicon vs. rules), processing correlates (memory vs. computation) and neuro-anatomical localizations (prefrontal vs. temporo-parietal perisylvian areas of the left hemisphere). In the present chapter, we explore the empirical and theoretical consequences of a distributed, integrative model of the mental lexicon, whereby words are seen as emergent properties of the functional interaction between basic, language-independent processing principles and the language- specific nature and organization of the input. From this perspective, language learning appears to be inextricably related to the way language is processed and internalized by the speakers, and key to an interdisciplinary understanding of such a way, in line with Tomaso Poggio's suggestion that the development of a cognitive skill is causally and ontogenetically prior to its execution (and sits "on top of it"). In particular, we discuss conditions, potential and prospects of the epistemological continuity between psycholinguistic and computational modelling of word learning, and illustrate the yet largely untapped potential of their integration. We use David Marr's hierarchy to clarify the complementarity of the two viewpoints. Psycholinguistic models are informative about how speakers learn to use language (interfacing Marr's levels 1 and 2). When we move from the psycholinguistic analysis of the functional operations involved in language learning to an algorithmic description of how they are computed, computer simulations can help us explore the relation between speakers' behavior and general learning principles in more detail. In the end, psycho-computational models can be instrumental to bridge Marr's levels 2 and 3, bringing us closer to understanding the nature of word knowledge in the brain.}, KEYWORDS = {mental lexicon, word storage and processing, psycholinguistics, computational linguistics, connectionist models, discriminative learning}, PAGES = {23-82}, URL = {https://www.degruyter.com/view/book/9783110440577/10.1515/9783110440577-002.xml}, VOLUME = {337}, DOI = {10.1515/9783110440577-002}, PUBLISHER = {De Gruyter Saur (Berlin/Munich, DEU)}, ISBN = {9783110440577}, BOOKTITLE = {Word Knowledge and Word Usage}, EDITOR = {Pirrelli, V. and Plag, I. and Dressler, W. U.}, } @INPROCEEDINGS{FERRO_2020_INPROCEEDINGS_FGC_441873, AUTHOR = {Ferro, M. and Giulivi, S. and Cappa, C.}, TITLE = {The AEREST reading database}, YEAR = {2020}, ABSTRACT = {Aerest is a reading assessment protocol for the concurrent evaluation of a child's decoding and comprehension skills. Reading data complying with the Aerest protocol were automatically collected and structured with the ReadLet web-based platform in a pilot study, to form the Aerest Reading Database. The content, structure and potential of the database are described here, together with the main directions of current and future developments.}, KEYWORDS = {reading database, reading efficiency, decoding, comprehension, multimodal analysis}, PAGES = {1-6}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-85097912116\&origin=inward}, VOLUME = {2769}, PUBLISHER = {Accademia University Press (Torino, ITA)}, ISSN = {1613-0073}, ISBN = {9791280136282}, CONFERENCE_NAME = {7th Italian Conference on Computational Linguistics (CLIC-IT'20)}, CONFERENCE_PLACE = {Bologna, Italy}, CONFERENCE_DATE = {01-03/03/2021}, BOOKTITLE = {CEUR workshop proceedings}, } @INPROCEEDINGS{PIRRELLI_2020_INPROCEEDINGS_PCCDFGMNT_442758, AUTHOR = {Pirrelli, V. and Cappa, C. and Crepaldi, D. and Del Pinto, V. and Ferro, M. and Giulivi, S. and Marzi, C. and Nadalini, A. and Taxitari, L.}, TITLE = {Tracking the pace of reading with finger movements}, YEAR = {2020}, ABSTRACT = {Recent experimental evidence in visual perception analysis shows that eye and finger movements strongly correlate during scene exploration, at both individual and group levels. A familiar context which exploits this synergistic behaviour is when children learn to read, with the practice of finger-pointing to text as a support for their attention focus, directional movement and voice-print match. Using a tablet to display short texts, we collected evidence on the finger-pointing behaviour of 3rd-6th Italian graders engaged in both silent and oral reading. "Finger-tracking" data, sampled by the tablet and aligned with the text, made it possible to time a child's reading paceat word and sentence level. Results are shown to replicate established benchmarks in the reading literature, such as the difference in reading pace between age-matched typical and atypical readers as a function of word frequency and length, and neighbourhood entropy and Old20. Atypical readers show increasing difficulty with longer words, with a steeper time increment for word length > 6, integrating previous evidence. In addition, neighbourhood density plays a sparse facilitative role in atypical reading, with no significant interaction with neighbourhood entropy, pointing to a non trivial developmental interplay between sublexical reading and the richness of the Italian orthographic-phonological lexicon. Despite their different dynamics, optical and tactile strategies for text exploration prove to be highly congruent: this suggests that finger-tracking can be used as an ecological proxy for eye-tracking in reading assessment.}, KEYWORDS = {Reading, Finger tracking, Mental Lexicon, Word frequency, Word Length, Neighbourhood entropy}, PAGES = {1}, URL = {https://osf.io/hr62g/}, CONFERENCE_NAME = {Words in the World International Conference}, CONFERENCE_PLACE = {Montreal (Canada)}, CONFERENCE_DATE = {16-18/10/2020}, } @ARTICLE{MARZI_2019_ARTICLE_MFP_406277, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {A processing-oriented investigation of inflectional complexity}, YEAR = {2019}, ABSTRACT = {Due to the typological diversity of their inflectional processes, some languages are intuitively more difficult than other languages. Yet, finding a single measure to quantitatively assess the comparative complexity of an inflectional system proves an exceedingly difficult endeavor. In this paper we propose to investigate the issue from a processing-oriented standpoint, using data processed by a type of recurrent neural network to quantitatively model the dynamic of word processing and learning in different input conditions. We evaluate the relative complexity of a set of typologically different inflectional systems (Greek, Italian, Spanish, German, English and Standard Modern Arabic) by training a Temporal Self-Organizing Map (TSOM), a recurrent variant of Kohonen's Self-Organizing Maps, on a fixed set of verb forms from top-frequency verb paradigms, with no information about the morphosemantic and morphosyntactic content conveyed by the forms. After training, the behavior of each language-specific TSOM is assessed on different tasks, looking at self-organizing patterns of temporal connectivity and functional responses. Our simulations show that word processing is facilitated by maximally contrastive inflectional systems, where verb forms exhibit the earliest possible point of lexical discrimination. Conversely, word learning is favored by a maximally generalizable system, where forms are inferred from the smallest possible number of their paradigm companions. Based on evidence from the literature and our own data, we conjecture that the resulting balance is the outcome of the interaction between form frequency and morphological regularity. Big families of stem-sharing, regularly inflected forms are the productive core of an inflectional system. Such a core is easier to learn but slower to discriminate. In contrast, less predictable verb forms, based on alternating and possibly suppletive stems, are easier to process but are learned by rote. Inflection systems thus strike a balance between these conflicting processing and communicative requirements, while staying within tight learnability bounds, in line with Ackermann and Malouf's Low Conditional Entropy Conjecture. Our quantitative investigation supports a discriminative view of morphological inflection as a collective, emergent system, whose global self-organization rests on a surprisingly small handful of language-independent principles of word coactivation and competition.}, KEYWORDS = {Morphological complexity, Discriminative learning, Recurrent neural networks (RNNs), self-organization, emergence, processing uncertainty, stem-family size}, PAGES = {1-23}, URL = {https://www.frontiersin.org/articles/10.3389/fcomm.2019.00048/full}, VOLUME = {4}, DOI = {10.3389/fcomm.2019.00048}, PUBLISHER = {Frontiers Media (Lausanne, Svizzera)}, ISSN = {2297-900X}, JOURNAL = {Frontiers in communication}, } @ARTICLE{CARDILLO_2018_ARTICLE_CFMP_396348, AUTHOR = {Cardillo, F. A. and Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {Deep Learning of Inflection and the Cell-Filling Problem}, YEAR = {2018}, ABSTRACT = {Machine learning offers two basic strategies for morphology induction: lexical segmentation and surface word relation. The first approach assumes that words can be segmented into morphemes. Inferring a novel inflected form requires identification of morphemic constituents and a strategy for their recombination. The second approach dispenses with segmentation: lexical representations form part of a network of associatively related inflected forms. Production of a novel form consists in filling in one empty node in the network. Here, we present the results of a task of word inflection by a recurrent LSTM network that learns to fill in paradigm cells of incomplete verb paradigms. Although the task does not require morpheme segmentation, we show that accuracy in carrying out the inflection task is a function of the model's sensitivity to paradigm distribution and morphological structure.}, KEYWORDS = {Deep Learning, LSTM, Cell-Filling Problem}, PAGES = {57-75}, URL = {https://publications.cnr.it/doc/396348}, VOLUME = {4}, PUBLISHER = {aAccademia University Press, Torino (Italia)}, ISSN = {2499-4553}, JOURNAL = {Italian Journal of Computational Linguistics}, } @ARTICLE{FERRO_2018_ARTICLE_FMP_397012, AUTHOR = {Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {Discriminative word learning is sensitive to inflectional entropy}, YEAR = {2018}, ABSTRACT = {Psycholinguistic evidence based on inflectional and derivational word families has emphasised the combined role of Paradigm Entropy and Inflectional Entropy in human word processing. Although the way frequency distributions affect behavioural evidence is clear in broad outline, we still miss a clear algorithmic model of how such a complex interaction takes place and why. The main challenge is to understand how the local interaction of learning and processing principles in morphology can result in global effects that require knowledge of the overall distribution of stems and affixes in word families. We show that principles of discriminative learning can shed light on this issue. We simulate learning of verb inflection with a discriminative recurrent network of specialised processing units, whose level of temporal connectivity reflects the frequency distribution of input symbols in context. We analyse the temporal dynamic with which connection weights are adjusted during discriminative learning, to show that self-organised connections are optimally functional to word processing when the distribution of inflected forms in a paradigm (Paradigm Entropy) and the distribution of their inflectional affixes across paradigms (Inflectional Entropy) diverge minimally.}, KEYWORDS = {discriminative learning, word processing, recurrent neural networks, relative entropy}, PAGES = {307-327}, URL = {https://www.rivisteweb.it/doi/10.1418/91871}, VOLUME = {XVII}, DOI = {10.1418/91871}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @INPROCEEDINGS{FERRO_2018_INPROCEEDINGS_FCGMNCP_390504, AUTHOR = {Ferro, M. and Cappa, C. and Giulivi, S. and Marzi, C. and Nahli, O. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {ReadLet: Reading for Understanding}, YEAR = {2018}, ABSTRACT = {This paper focuses on motivation, objectives, design issues and preliminary results of ReadLet, an ICT platform for assessing reading efficiency in primary school children. Test data are discussed on a sample of 200 early graders, reading French, Italian and Standard Modern Arabic (SMA).}, KEYWORDS = {Reading, text comprehension, Specific Learning Disorders, multimodal signal processing, cloud computing, portable assistive technology}, PAGES = {404-409}, URL = {https://publications.cnr.it/doc/390504}, PUBLISHER = {IEEE (New York, USA)}, ISBN = {978-1-5386-4385-3}, CONFERENCE_NAME = {IEEE-CIST2018 LED-ICT}, CONFERENCE_PLACE = {Marrakech, Morocco}, CONFERENCE_DATE = {21-27/10/2018}, } @INPROCEEDINGS{MARZI_2018_INPROCEEDINGS_MFNBBP_388016, AUTHOR = {Marzi, C. and Ferro, M. and Nahli, O. and Belik, P. and Bompolas, S. and Pirrelli, V.}, TITLE = {Evaluating Inflectional Complexity Crosslinguistically: a Processing Perspective}, YEAR = {2018}, ABSTRACT = {The paper provides a cognitively motivated method for evaluating the inflectional complexity of a language, based on a sample of "raw" inflected word forms processed and learned by a recurrent self-organising neural network with fixed parameter setting. Training items contain no information about either morphological content or structure. This makes the proposed method independent of both meta-linguistic issues (e.g. format and expressive power of descriptive rules, manual or automated segmentation of input forms, number of inflectional classes etc.) and language-specific typological aspects (e.g. word-based, stem-based or template-based morphology). Results are illustrated by contrasting Arabic, English, German, Greek, Italian and Spanish.}, KEYWORDS = {paradigm-based morphology, inflectional complexity, prediction-based processing, recurrent self-organising networks, Statistical And Machine Learning Methods, Language Modelling}, PAGES = {3860-3866}, URL = {http://www.lrec-conf.org/proceedings/lrec2018/summaries/745.html}, VOLUME = {2018}, PUBLISHER = {European language resources association (ELRA) (Paris, FRA)}, ISBN = {979-10-95546-00-9}, CONFERENCE_NAME = {Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, CONFERENCE_PLACE = {Miyazaki, Japan}, CONFERENCE_DATE = {7-12/05/2018}, BOOKTITLE = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, EDITOR = {Calzolari, N. and Choukri, K. and Cieri, C. and Declerck, T. and Goggi, S. and Hasida, K. and Isahara, H. and Maegaard, B. and Mariani, J. and Mazo, H. and Moreno, A. and Odijk, J. and Piperidis, S. and Tokunaga, T.}, } @INPROCEEDINGS{CAPPA_2018_INPROCEEDINGS_CFGMNCP_396593, AUTHOR = {Cappa, C. and Ferro, M. and Giulivi, S. and Marzi, C. and Nahli, O. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {ReadLet: piattaforma ICT per valutare l'efficienza di lettura}, YEAR = {2018}, ABSTRACT = {ReadLet è una piattaforma ICT pensata per valutare accuratamente l'efficienza di lettura nei bambini della scuola primaria. Combina tecnologia ICT portatile e cloud-computing con una serie di moduli software, specifici per modalità di somministrazione. Questi, implementati come servizi web, includono: i) valutazione dell'elaborazione del testo e della leggibilità; ii) valutazione della velocità di lettura (ad alta voce e silente) e delle sue fluttuazioni); iii) valutazione della correttezza della decodifica ad alta voce; iv) valutazione della comprensione del testo (in lettura silente e da ascolto). Un prototipo della tecnologia ReadLet è stato sperimentato su circa 200 alunni (8-11 anni), che variano per stato socio-economico, lingua (italiana, francese, araba) e area geografica (Italia, Svizzera, Marocco). L'utilizzo del tablet per la lettura è stato percepito dai bambini come un'esperienza coinvolgente e piacevole. Gli insegnanti hanno trovato lo strumento facile da utilizzare e in grado di fornire maggiori informazioni rispetto agli strumenti tradizionali.}, KEYWORDS = {leggere per capire, disturbi del linguaggio, screening}, URL = {https://www.airipa.it/congresso/pluginfile.php/2781/mod_resource/content/1/Programma%20Congresso%20AIRIPA_Arezzo_dettagliato-3.pdf}, CONFERENCE_NAME = {XXVII Congresso Nazionale AIRIPA}, CONFERENCE_PLACE = {Arezzo (Italy)}, CONFERENCE_DATE = {28-29/09/2018}, } @INPROCEEDINGS{FERRO_2018_INPROCEEDINGS_FCGMCP_396591, AUTHOR = {Ferro, M. and Cappa, C. and Giulivi, S. and Marzi, C. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {ReadLet: an ICT platform for the assessment of reading efficiency in early graders}, YEAR = {2018}, ABSTRACT = {Reading is not just word decoding, but the joint product of decoding and deep linguistic comprehension [ 1 , 2 ]. Effective linguistic comprehension relies on language skills such as semantic and syntactic awareness. Both decoding and linguistic comprehension are necessary for reading comprehension, and neither is by itself sufficient [ 2 ]. However, current protocols for reading assessment measure decoding (reading accuracy and speed) and reading comprehension separately [ 3 , 4 , 5 ]. This does not allow evaluation of reading efficiency [ 6 ], defined as the ability to fully understand connected texts by minimising reading time, a cognitive ability that lies at the roots of students' academic achievement [ 8 , 7 ]. ReadLet is an ICT platform specifically designed to provide accurate, evidence-based assessment of reading efficiency in early grade children, by offering an ecological, non-invasive protocol for extensive data elicitation, storage and analysis. With ReadLet, early graders at school can read a one or two page text displayed on a tablet touchscreen, either silently or aloud. Children are asked to slide their finger across the words as they read, to guide directional tracking. After reading, the child is prompted with a few multiple-answer questions on text content presented one at a time, while the text remains displayed on the screen for the child to be able to retrieve relevant information. In the process, the tablet keeps track of time-aligned multimodal data: voice recording, finger sliding time, time of reading, time of question answering, and number of correct answers. Data are recorded, stored locally, sent to the ReadLet server through an internet connection, and processed remotely by a battery of cloud-based services, analysing data automatically to produce a detailed quantitative signature of each reading session. A server-based database aggregates anonymised data to make them available for specialists. Also individual's longitudinal profiles are stored, for them be queried and inspected upon authorised access. The platform combines portable ICT technology and cloud computing with a number of modality-specific software modules, implemented as web services including: i) a text processing and readability assessment service, consisting in a battery of tools for automated linguistic annotation of written texts and a machine-learning component assigning a readability score to annotated texts [ 9 ]; ii) a finger touch processing service aligning the child's finger sliding with the written text and measuring speed fluctuations; iii) a speech processing and decoding assessment service, aligning the acoustic record of child's reading with the written text and assessing correctness of recoding [ 10 ]. At the time of writing, the platform includes the first two modules only. Preliminary testing of a prototype version of ReadLet technology with a population of about 200 pupils aged 8 to 11, both male and female, varying for socio-economic status, language (Italian, French and Arabic) and geographical area (Italy and Morocco), showed that children are extremely responsive to using a tablet for reading, and very easy to engage in what they perceive as an enjoyable experience. We expect online databases of automatically classified cross-sectional and longitudinal data, accurate statistical modelling and developmental trends of reading literacy to help education professionals and clinical specialists assess the level of reading skills reached by the child, and decide which intervention programmes and measures are most appropriate. While information technology cannot and should not supplant the role and professional judgement of teachers and therapists, the project intends to provide portable tools, models and data for timely screening and daily management of reading difficulties and disorders.}, KEYWORDS = {reading efficiency, decoding, comprehension, language specific disorders}, PAGES = {61-61}, URL = {https://mentallexicon2018.ca/}, CONFERENCE_NAME = {11th International Conference on the Mental Lexicon}, CONFERENCE_PLACE = {Edmonton, Alberta (Canada)}, CONFERENCE_DATE = {25-28/09/2018}, } @INPROCEEDINGS{LEONI_2018_INPROCEEDINGS_LMCFG_396592, AUTHOR = {Leoni, F. and Muzio, C. and Cappa, C. and Ferro, M. and Giulivi, S.}, TITLE = {Il progetto AEREST: primi risultati in Italia e in Canton Ticino}, YEAR = {2018}, ABSTRACT = {Il progetto AEREST, per una valutazione ecologica dell'efficienza di lettura, è attualmente in corso presso alcune classi di scuola primaria di istituti italiani e ticinesi. Si presentano qui i risultati ottenuti a seguito della prima sessione di raccolta dati, che si è svolta nell'A.A. 2017-18 su circa 160 bambini italofoni di età compresa tra 8 e 11 anni. Lo scopo di questa prima fase sperimentale è stato duplice: 1. ottenere indicazioni sull'efficacia, ai fini della valutazione dell'efficienza di lettura, dei testi utilizzati nelle prove di cui si compone il test AEREST; 2. ottenere indicazioni sulla fattibilità dell'implementazione dello screening su tablet, in termini di facilità di somministrazione e di gradimento da parte dei soggetti; 3. esplorare e confrontare le performance di lettura nel campione italiano e ticinese, al fine di individuare strategie didattiche volte a potenziare le eventuali abilità carenti.}, KEYWORDS = {efficienza di lettura, screening}, URL = {https://www.airipa.it/congresso/pluginfile.php/2781/mod_resource/content/1/Programma%20Congresso%20AIRIPA_Arezzo_dettagliato-3.pdf}, CONFERENCE_NAME = {XXVII Congresso Nazionale AIRIPA}, CONFERENCE_PLACE = {Arezzo (Italy)}, CONFERENCE_DATE = {28-29/09/2018}, } @INPROCEEDINGS{MARZI_2018_INPROCEEDINGS_MFP_396356, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Is inflectional irregularity dysfunctional to human processing?}, YEAR = {2018}, ABSTRACT = {Regularly inflected verb forms are classically associated with the formal transparency and predictability of their internal constituents [ 1 , 2 , 3 ]. Transparency ensures that full forms can be segmented uniquely into their internal constituents: as in walk-s/walk-ed. Predictability allows for a speaker to fill in an empty paradigm cell, using information from other known forms of the same lexical paradigm and its inflection macro-class. From this perspective, irregulars appear to be dysfunctional to the human processing system, as they make it hard to infer - say - bought from buy , or segment bought appropriately into its constituent parts. Likewise, an influential psycholinguistic tradition relegates irregulars to the lexical store, whereas regulars are segmented by rules into their simpler constituents [ 4 , 5 ]. Here, we offer a few reasons for questioning this view. First, transparency and predictability are not dichotomous notions. Secondly, their influence on processing is not unidirectional. Unpredictable stems in irregularly inflected forms of complex inflectional systems provide a lot of processing information, by dynamically constraining the number of possible alternative endings during serial processing. Thirdly, acquisition of word inflection does not consist in associating co-occurring cues and outcomes, but in discriminating between multiple cues that are constantly in competition for their predictive value for a given outcome. We present the results of a few computer simulations with Self-organising Recurrent Neural Networks (TSOMs, [ 8 , 9 ]) that learn how to inflect high-frequency verb paradigms in 6 languages: English, German, Italian, Modern Greek, Modern Standard Arabic and Spanish. After training, each TSOM was tested on a word recognition (serial recoding) and a word production (serial recall) task, and results were analysed with generalised regression models. Processing uncertainty is differently apportioned on regulars and irregulars, depending on the nature of the processing task. While irregulars are harder to produce when they are unknown because they typically have fewer neighbours than regulars have, they are readily accessed once they are acquired, for exactly the same reason. Our data are in line with psycholinguistic evidence [ 10 , 11 ] that lexical processing is paced by two types of uniqueness point: Marslen-Wilson's Uniqueness Point (UP), distinguishing unrelated onset-overlapping words [ 12 ], and the Complex Uniqueness Point (CUP), distinguishing paradigmatically-related words [ 11 ]. Late UPs are inhibitory and elicit prolonged reaction times in acoustic word recognition, explaining an early delay in word recognition of irregular stems. Similarly, late CUPs are inhibitory, and this accounts for a slowdown in the processing advantage of regulars, compared to irregulars, after UP. These structural factors interact in a variety of ways and concurrently affect human processing, to show that irregularly-inflected forms may in fact reflect communicative and processing constraints of the word processor. They provide strong evidence against a processing architecture that assumes compartmentalized, independent processing routes for some specific combinations of these factors (e.g. a rule-based route for a combination of transparency and predictability, and a memory-based route for all other combinations). In addition, they seem incompatible with Bayesian approaches to auditory word comprehension ignoring a word's internal structure [ 13 ]. We suggest that a different design of the human language processor, based on a computational architecture integrating memory and processing as two different dynamics of the same underlying mechanism, can shed light on the complexity of inflection, and vindicate the role of irregular inflection in the system.}, KEYWORDS = {inflectional processing, temporal self organizing maps, letter prediction, morpheme boundary}, PAGES = {60-60}, URL = {https://mentallexicon2018.ca/}, CONFERENCE_NAME = {11th International Conference on the Mental Lexicon}, CONFERENCE_PLACE = {Edmonton, Alberta (Canada)}, CONFERENCE_DATE = {25-28/09/2018}, } @INPROCEEDINGS{PIRRELLI_2018_INPROCEEDINGS_PFMGSM_396353, AUTHOR = {Pirrelli, V. and Ferro, M. and Marzi, C. and Gagné, C. and Spalding, T. and Marelli, M.}, TITLE = {Processing compounds: what frequency (alone) cannot explain}, YEAR = {2018}, ABSTRACT = {Observed elevation in typing latency for the initial letter of the second constituent of an English compound, compared with the typing time of the final letter of the first constituent (Gagné \& Spalding 2016), suggests that both compounds ( snowball ) and pseudo-compounds ( carpet ) are decomposed but also that full form representations are available in the lexical store. To gain further insight into the lexical representations underlying typing, we used computational modelling. In particular, we used superpositional models of word memory, based on Self-Organising Recurrent Maps (TSOMs) (Ferro et al. 2016; Marzi et al. 2016), where both simple and compound words are processed (and stored) using the same pool of processing (and memory) resources, to model the elevation in typing time at the constituent boundary and the rate of typing. In addition, we also considered models based in the Compositional Distributional Semantics framework (CAOSS, Marelli et al. 2017), to simulate independent effects of semantic transparency on compound typing (Gagné \& Spalding 2016). Due to co-activation and competition between compounds and their constituent words in TSOMs, levels of activation of processing nodes per letter positions appear to reflect degrees of context-sensitive predictability: the higher the level, the more expected the letter in that position. In English compounds, activation levels appeared to exhibit a characteristically U-shaped pattern, with min values centred on the constituent boundary. A similar pattern was found for pseudo-compounds, which nonetheless present a less pronounced U-shaped pattern and a higher activation value at the morpheme boundary than compounds do. The difference is in line with the higher speed-up rate in typing pseudo-compounds than compounds reported in Gagné and Spalding (2016). TSOMs were trained on letter-based representations, so computer experiments could simulate peripheral effects of serial processing of compound structure before lexical access. To investigate post-lexical issues, we also tested computational models of generation of the meanings of novel compounds based on CAOSS, which proved to be able to account for well-established relational effects in compound processing (Gagné 2001; Gagné \& Shoben 1997) with an unsupervised data-driven framework (Marelli et al. 2017). We ran a mixed-effects regression analysis of the data in Gagné and Spalding (2016) using vector-semantics estimates and TSOM activation levels to predict typing time for the initial letter of the second constituent. There was a negative effect of TSOM letter activation levels: i.e. the more active a letter node is, the faster a subject is at typing the letter ( t =-2.7 p =.007). Also, there was a positive effect of CAOSS-based compositionality estimates: i.e. the more easily a compound's lexicalized meaning can be obtained through compositional operations on single constituent vectors, the slower participants were at typing the first letter of the second constituent ( t =2.4, p =.017). These results have interesting implications for an integrative computational architecture accounting for the whole range of experimental evidence reported by Gagné and Spalding (2016). In particular we will focus on evidence of a stronger competition (and longer typing time) in Transparent-Transparent and Transparent-Opaque compounds, vs. Opaque-Transparent compounds, which gives an indication of a non-trivial interaction between semantic compositionality and serial processing effects.}, KEYWORDS = {compound processing, Temporal Self-organizing Map, letter production latency, constituent boundary}, PAGES = {60-60}, URL = {https://mentallexicon2018.ca/}, CONFERENCE_NAME = {11th International Conference on the Mental Lexicon}, CONFERENCE_PLACE = {Edmonton (Canada)}, CONFERENCE_DATE = {25-28/09/2018}, } @ARTICLE{BOMPOLAS_2017_ARTICLE_BFMCP_380237, AUTHOR = {Bompolas, S. and Ferro, M. and Marzi, C. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {For a performance-oriented notion of regularity in inflection: the case of Modern Greek conjugation}, YEAR = {2017}, ABSTRACT = {Paradigm-based approaches to word processing/learning assume that word forms are not acquired in isolation, but through associative relations linking members of the same word family (e.g. a paradigm, or a set of forms filling the same paradigm cell). Principles of correlative learning offer a set of equations that are key to modelling this complex dynamic at a considerable level of detail. We use these equations to simulate acquisition of Modern Greek conjugation, and we compare the results with evidence from German and Italian. Simulations show that different Greek verb classes are processed and acquired differentially, as a function of their degrees of formal transparency and predictability. We relate these results to psycholinguistic evidence of Modern Greek word processing, and interpret our findings as supporting a view of the mental lexicon as an emergent integrative system.}, KEYWORDS = {paradigm-based morphology, gradient (ir)regularity, recurrent self-organisng networks}, PAGES = {77-92}, URL = {http://www.ai-lc.it/IJCoL/v3n1/IJCOL_3_1_5_bompolas_et_al.pdf?v=2a47ad90f2ae}, VOLUME = {3}, PUBLISHER = {aAccademia University Press, Torino (Italia)}, ISSN = {2499-4553}, JOURNAL = {Italian Journal of Computational Linguistics}, } @ARTICLE{MARZI_2017_ARTICLE_MFN_363116, AUTHOR = {Marzi, C. and Ferro, M. and Nahli, O.}, TITLE = {Arabic word processing and morphology induction through adaptive memory self-organisation strategies}, YEAR = {2017}, ABSTRACT = {Aim of the present study is to model the human mental lexicon, by focussing on storage and processing dynamics, as lexical organisation relies on the process of input recoding and adaptive strategies for long-term memory organisation. A fundamental issue in word processing is represented by the emergence of the morphological organisation level in the lexicon, based on paradigmatic relations between fully-stored word forms. Morphology induction can be defined as the task of perceiving and identifying morphological formatives within morphologically complex word forms, as a function of the dynamic interaction between lexical representations and distribution and degrees of regularity in lexical data. In the computational framework we propose here (TSOMs), based on Self-Organising Maps with Hebbian connections defined over a temporal layer, the identification/perception of surface morphological relations involves the alignment of recoded representations of morphologically-related input words. Facing a non-concatenative morphology such as the Arabic inflectional system prompts a reappraisal of morphology induction through adaptive organisation strategies, which affect both lexical representations and long-term storage. We will show how a strongly adaptive self-organisation during training is conducive to emergent relations between word forms, which are concurrently, redundantly and competitively stored in human mental lexicon, and to generalising knowledge of stored words to unknown forms.}, KEYWORDS = {Non-concatenative morphological structure, Lexical storage and access, Topological alignment, Synchronisation, Self-Organising Maps}, PAGES = {179-188}, URL = {http://www.sciencedirect.com/science/article/pii/S1319157816301148}, VOLUME = {29}, DOI = {10.1016/j.jksuci.2016.11.006}, PUBLISHER = {Elsevier (Amsterdam, Paesi Bassi)}, ISSN = {2213-1248}, JOURNAL = {Journal of King Saud University. Computer and information sciences (Online)}, } @INPROCEEDINGS{CARDILLO_2017_INPROCEEDINGS_CFMP_381090, AUTHOR = {Cardillo, F. A. and Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {How "deep" is learning word inflection?}, YEAR = {2017}, ABSTRACT = {Machine learning offers two basic strategies for morphology induction: lexical segmentation and surface word relation. The first one assumes that words can be segmented into morphemes. Inducing a novel inflected form requires identification of morphemic constituents and a strategy for their recombination. The second approach dispenses with segmentation: lexical representations form part of a network of associatively related inflected forms. Production of a novel form consists in filling in one empty node in the network. Here, we present the results of a recurrent LSTM network that learns to fill in paradigm cells of incomplete verb paradigms. Although the process is not based on morpheme segmentation, the model shows sensitivity to stem selection and stem-ending boundaries.}, KEYWORDS = {LSTM, Morphology induction, Cognitive modelling}, PAGES = {77-82}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-85037368972\&origin=inward}, VOLUME = {2006}, DOI = {10.4000/books.aaccademia.2314}, PUBLISHER = {Accademia University Press (Torino, ITA)}, ISSN = {1613-0073}, ISBN = {978-88-99982-76-8}, CONFERENCE_NAME = {Fourth Italian Conference on Computational Linguistics}, CONFERENCE_PLACE = {Roma}, CONFERENCE_DATE = {11-13/12/2017}, BOOKTITLE = {Proceedings of the Fourth Italian Conference on Computational Linguistics (CLiC-it 2017)}, EDITOR = {Basili, R. and Nissim, M. and Satta, G.}, } @INPROCEEDINGS{BOMPOLAS_2017_INPROCEEDINGS_BMFCPR_381125, AUTHOR = {Bompolas, S. and Marzi, C. and Ferro, M. and Cardillo, F. A. and Pirrelli, V. and Ralli, A.}, TITLE = {Transparency and predictability in Modern Greek conjugation: Implications for models of word processing}, YEAR = {2017}, ABSTRACT = {We argue that the Greek evidence calls for a substantial revision of the clear-cut interaction between transparency/predictability and regularity, to make room for a more process-oriented notion of regularity. According to this view, regularity is no longer an epiphenomenon of the design of the human language faculty and the purported dualism between rule-based and memory-based routes, but the graded result of the varying interaction of several structural factors concurrently affecting the human word processor.}, KEYWORDS = {Inflectional regularity, Word Processing, Modern Greek Conjugation}, PAGES = {17-19}, URL = {http://www.lilec.it/mmm/wp/wp-content/uploads/2017/02/Book-of-abstracts_MMM11_Final.pdf}, CONFERENCE_NAME = {MMM 11: 11th Mediterranean Morphology Meeting}, CONFERENCE_PLACE = {Cyprus}, CONFERENCE_DATE = {22-25/06/2017}, } @INPROCEEDINGS{PIRRELLI_2017_INPROCEEDINGS_PMFC_381117, AUTHOR = {Pirrelli, V. and Marzi, C. and Ferro, M. and Cardillo, F. A.}, TITLE = {Paradigm Relative Entropy and Discriminative Learning}, YEAR = {2017}, ABSTRACT = {In the present contribution, we show that principles of discriminative learning of symbolic time series go a long way in accounting for these effects, thus making an important contribution to our understanding of the human lexical processor and its sensitivity to word distributions both within and across paradigms.}, KEYWORDS = {Paradigm Entropy, Discriminative Learning, Mental Lexicon, Verb Inflection}, PAGES = {5}, URL = {http://w3.erss.univ-tlse2.fr/ParadigMo2017/program.html}, CONFERENCE_NAME = {ParadigMo 2017: First Workshop on Paradigmatic Word Formation Modeling}, CONFERENCE_PLACE = {Toulouse}, CONFERENCE_DATE = {19-20/06/2017}, } @ARTICLE{MARZI_2016_ARTICLE_MFCP_360723, AUTHOR = {Marzi, C. and Ferro, M. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {Effects of frequency and regularity in an integrative model of word storage and processing}, YEAR = {2016}, ABSTRACT = {Considerable evidence has accrued on the role of paradigms as both theoretical and cognitive structures regimenting the way words are processed and acquired. The evidence supports a view of the lexicon as an emergent integrative system, where word forms are concurrently and competitively stored as repeatedly successful processing patterns, and on-line processing crucially depends on the internal organisation of stored patterns.}, KEYWORDS = {Lexical access, word recall, serial processing, parallel activation, inflectional paradigms, mental lexicon}, PAGES = {79-114}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-84986550295\&origin=inward}, VOLUME = {28}, PUBLISHER = {Pacini (Ospedaletto, Italia)}, ISSN = {1120-2726}, JOURNAL = {Rivista di Linguistica}, } @INPROCEEDINGS{BOMPOLAS_2016_INPROCEEDINGS_BMFCP_362297, AUTHOR = {Bompolas, S. and Marzi, C. and Ferro, M. and Cardillo, F. A. and Pirrelli, V.}, TITLE = {Reassessing inflectional regularity in Modern Greek conjugation}, YEAR = {2016}, ABSTRACT = {Paradigm-based approaches to word processing/learning assume that word forms are not acquired in isolation, but through associative relations linking members of the same word family (e.g. a paradigm, or a set of forms filling the same paradigm cell). Principles of correlative learning offer a set of dynamic equations that are key to modelling this complex dynamic at a considerable level of detail. We use these dynamic equations to simulate acquisition of Modern Greek conjugation, and we compare the results with evidence from German and Italian. Simulations show that different Greek verb classes are processed and acquired differentially, depending on their degrees of formal transparency and predictability. We relate these results to psycholinguistic evidence on Modern Greek word processing, and interpret our findings as supporting a view of the mental lexicon as an emergent integrative system.}, KEYWORDS = {word processing, paradigm-based learning, morphological processing, Greek stem allomoprhy, Temporal Self-Organising Map}, PAGES = {72-77}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-85009242702\&origin=inward}, VOLUME = {1749}, DOI = {10.4000/books.aaccademia.1721}, PUBLISHER = {Accademia University Press (Torino, ITA)}, ISSN = {1613-0073}, ISBN = {978-88-99982-08-9}, CONFERENCE_NAME = {Third Italian Conference on Computational Linguistics (CLiC-it 2016) \& Fifth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian. Final Workshop (EVALITA 2016)}, CONFERENCE_PLACE = {Napoli, Italy}, CONFERENCE_DATE = {05-07/12/2016}, BOOKTITLE = {CLiC-it \& EVALITA 2016-Proceedings of Third Italian Conference on Computational Linguistics (CLiC-it 2016) \& Fifth Evaluation Campaign of Natural Language Processing and Speech Tools for Italian. Final Workshop (EVALITA 2016)}, EDITOR = {Basile, P. and Corazza, A. and Monetmagni, S. and Nissim, M. and Patti, V. and Semeraro, G. and Sprugnoli, R.}, } @INPROCEEDINGS{FERRO_2016_INPROCEEDINGS_FCPGS_362349, AUTHOR = {Ferro, M. and Cardillo, F. A. and Pirrelli, V. and Gagné, C. L. and Spalding, T. L.}, TITLE = {Written word production and lexical self-organisation: evidence from English (pseudo)compounds}, YEAR = {2016}, ABSTRACT = {Elevation in typing latency for the initial letter of the second constituent of an English compound, relative to the latency for the final letter of the first constituent of the same compound, provides evidence that implementation of a motor plan for written compound production involves smaller constituents, in both semantically transparent and semantically opaque compounds. We investigate here the implications of this evidence for algorithmic models of lexical organisation, to show that effects of differential perception of the internal structure of compounds and pseudo-compounds can also be simulated as peripheral stages of lexical access by a self-organising connectionist architecture, even in the absence of morphosemantic information. This complementary evidence supports a maximizationof-opportunity approach to lexical modelling, accounting for the integration of effects of pre-lexical and lexical access.}, KEYWORDS = {compound, pseudo-compound, written word production, lexical self-organisation, temporal self organising map}, PAGES = {146-151}, URL = {http://ceur-ws.org/Vol-1749/}, VOLUME = {1749}, DOI = {10.4000/books.aaccademia.1775}, PUBLISHER = {Accademia University Press (Torino, ITA)}, ISSN = {1613-0073}, ISBN = {9788899982546}, CONFERENCE_NAME = {Third Italian Conference on Computational Linguistics (CLiC-it 2016)}, CONFERENCE_PLACE = {Napoli (Italia)}, CONFERENCE_DATE = {5-6/12/2016}, BOOKTITLE = {Proceedings CLiC-it 2016}, EDITOR = {Basile, P. and Corazza, A. and Cutugno, F. and Montemagni, S. and Nissim, M. and Patti, V. and Semeraro, G. and Sprugnoli, R.}, } @INPROCEEDINGS{RECCHIA_2016_INPROCEEDINGS_RFMD_362391, AUTHOR = {Recchia, V. and Ferro, M. and Maglie, R. and Dodaro, A.}, TITLE = {Readability of current patient information leaflets for informed consent in UK radiotherapy centers}, YEAR = {2016}, ABSTRACT = {Background: Guidelines on informed consent recommend the use of plain language and readability standards to enhance patient's comprehension, engagement and shared decision making. Aim: To assess the readability of current patient information leaflets (PILs) used for informed consent in radiotherapy. Methods: We evaluated PILs (n=38) from three radiation therapy centers in UK. They regard the most common radiation therapy techniques for different kinds of cancer and body disctricts, such as bladder, bowel, colo-rectum, brain, breast-chest, femal pelvis, prostate, lung, linphomas, stomach. We analyzed each text with Flesch-Kincaid (F-K) grade level, with higher numbers indicating harder-to-read text (from 0 = easy, to 25 = difficult). Then, we compared the related grade levels to the health literacy recommended standard of US grade level 5, indicating that patient education texts might be understood by a typical student in the US primary school. Results: Readibility is suboptimal for the analised PILs (red, green and blue points in the figure) and should be improved with respect to the international standard score (red dotted line in the figure). The results show a mean grade level equal to 8.1 (std = 0.8), thus suggesting the need of a 3-points decrease on average. Conclusion: Current PILs for informed consent in the three analised radiotherapy centers are hardly readable for the average patient. Although the readability scores achieved in the three centers is not very low, substantially higher readability scores should be achieved with novel PILs which explicitly discuss risks/benefits and other elements relevant for informed consent, and should be prepared by following standard recommendations of plain language.}, KEYWORDS = {Ethics and communication, Communicating Risk and Uncertainty, Health Literacy}, PAGES = {1}, URL = {http://www.communication.aau.dk/research/dihm/events/comet2016/}, CONFERENCE_NAME = {COMMUNICATION, MEDICINE AND ETHICS CONFERENCE 2016}, CONFERENCE_PLACE = {Aalborg, Denmark}, CONFERENCE_DATE = {4-6/6/2016}, } @INCOLLECTION{PIRRELLI_2015_INCOLLECTION_PFM_330234, AUTHOR = {Pirrelli, V. and Ferro, M. and Marzi, C.}, TITLE = {Computational complexity of abstractive morphology}, YEAR = {2015}, ABSTRACT = {Abstractive and constructive approaches to word structure make radically different assumptions concerning nature and role of the building blocks that make up a speaker's morphological competence. In this contribution, we show that the two views are also computationally different. In particular, we contend that a number of problems arising in connection with a subsymbolic implementation of the constructive view (as epitomised by classical multi-layered perceptrons) are tackled effectively, or disappear altogether, in a neurally-inspired implementation of associative networks, resting on key-notions such as self-organization and emergence. A particular variant of Kohonen's Self-Organizing Map is introduced as a model to explore and assess the implications of an abstractive approach in terms of its computational complexity. Details of the model (Temporal Self-Organizing Map, TSOM) and experimental data are shown to illustrate the interplay between processing and storage in language acquisition.}, KEYWORDS = {Word processing, computational complexity, mental lexicon, dynamic memories, self-organisation, word structure, morphology}, PAGES = {141-166}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-84938781714\&origin=inward}, DOI = {10.1093/acprof:oso/9780198723769.003.0008}, PUBLISHER = {Oxford University Press (Oxford, GBR)}, ISBN = {978-0-19-872376-9}, BOOKTITLE = {Understanding and Measuring Mprphological Complexity}, EDITOR = {Baerman, M. and Brown, D. and Corbett, G. G.}, } @EDITORIAL{PIRRELLI_2015_EDITORIAL_PMF_329357, AUTHOR = {Pirrelli, V. and Marzi, C. and Ferro, M.}, TITLE = {Proceedings of the NetWordS Final Conference on Word Knowledge and Word Usage: Representations and Processes in the Mental Lexicon}, YEAR = {2015}, ABSTRACT = {The international conference "Word Knowledge and Word Usage: Representations and processes in the mental lexicon" is the final outcome of 4 years of intense multi-disciplinary research networking and cooperation funded by the European Science Foundation within the framework of the NetWordS programme (May 2011 - April 2015). NetWordS' mission was to bring together experts of various research fields (from brain sciences and computing to cognition and linguistics) and of different theoretical inclinations, to advance the current awareness of theoretical, typological, psycholinguistic, computational and neurophysiological evidence on the structure and processing of words, with a view to developing novel research paradigms and bringing up a new generation of language scholars. The conference was intended to provide a first forum for assessing current progress of crossdisciplinary research on language architecture and usage, and discussing prospects of future synergy. People are known to memorise, parse and access words in a context-sensitive and opportunistic way, by caching their most habitual and productive processing patterns into routinized behavioural schemes. Speakers not only take advantage of token-based information such as frequency of individual, holistically stored words, but they are also able to organise stored words through paradigmatic structures (or word families) whose overall size and frequency is an important determinant of ease of lexical access and interpretation. Accordingly, lexical organisation is not necessarily functional to descriptive economy and minimisation of storage, but to more performance-oriented factors such as efficiency of memorisation, access and recall. Usage-based approaches to word processing lend support to this view, to promote explanatory frameworks that aim to investigate the stable correlation patterns linking distributional entrenchment of lexical units with productivity, internal structure and ease of interpretation. Ultimately, this is intended to establish a deep interconnection between performance-oriented,low-level lexical functions such as memorisation, rehearsal, access and recall, and their neuroanatomical correlates.}, KEYWORDS = {mental lexicon, linguistics, brain sciences, psycholinguistics, computing, cognition}, PAGES = {1-189}, URL = {http://ceur-ws.org/Vol-1347/}, VOLUME = {1347}, PUBLISHER = {CEUR-WS. org (Aachen, DEU)}, } @INPROCEEDINGS{FERRO_2015_INPROCEEDINGS_FMP_331183, AUTHOR = {Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {Lexical parsability and morphological structure}, YEAR = {2015}, ABSTRACT = {A classical tenet in the psycholinguistic literature on the mental lexicon is that a parsed affix presents high activation levels (and thus contributes to activation spreading to other words with the same affix), and that such levels are tightly correlated with the affix productivity. In a number of influential papers, it has been suggested that parsability criteria interact with frequency to define morphological productivity in the lexicon. For example, the frequency of a derivative (e.g. government) relative to its base (govern) is shown to be a good predictor for parsability/productivity. The higher the frequency ratio, the more likely the morphological structure to be perceived, and the associated affix to be used productively. The present contribution intends to offer a computational explanatory basis for this correlational evidence, and assess its applicability to the acquisition of complex inflectional paradigms. In those languages, like Italian and German, whose inflection is stem-based rather than word-based, there is often no single paradigmatic form which can act as a base by being properly contained in all other inflected variants. Yet, it seems intuitive to suggest that verbs that are inflected for one paradigm cell only (e.g. neighbouring), are learned earlier and more easily but exhibit lower levels of perceived inflectional structure than verbs with richer paradigms. This appears to be in good accord with experimental evidence of time latencies in lexical decision, which are shown to correlate negatively with token frequency, paradigm size and paradigm entropy. Our simulations, based on Temporal Self-Organizing Maps (TSOMs) allow us to establish an interesting connection between inflectional parsability, frequency-based paradigm structure, and acquisitional constraints on the interaction between the human processor and working memory. Self-organising topological models of the mental lexicon can mimic the spatial and temporal organization of memory structures supporting the processing of symbolic sequences, and can provide an interesting framework for testing integrative accounts of lexical processing/acquisition as the complex result of general-purpose operations on word stimuli (e.g. working memory, long-term storage, sensory-motor mapping, rehearsal, unit integration, unit analysis, executive control, time-series processing), in line with recent acquisitions on the neuro-functional architecture of the perisylvian language network in the left hemisphere of human brain. Simulations of the incremental acquisition of "mini-paradigms" (small islands of morphological contrast encompassing up to three different forms for the same verb support the hypothesis that perception of structure (parsability) and morphological productivity strongly correlate in the inflectional lexica of German and Italian. In particular, by monitoring longitudinal progress in storage and generalisation of differently distributed inflectional paradigms in the two languages, we show that: i) high-frequency forms are stored and accessed significantly earlier than low-frequency forms; ii) deeply entrenched but paradigmatically isolated forms tend to block usage of other forms in the same paradigm; iii) low-frequency evenly distributed (highly entropic) intra-paradigmatic forms are acquired later but are easily extended. Our investigation credits the proposed computational framework with psycholinguistic plausibility, and grounds parsability-based models of morphological productivity on a specific, explicit proposal of lexical architecture. This provides an explanatory basis for both psycholinguistic and linguistic accounts of morphological structure, and offers an intermediate framework for scientific inquiry bridging the gap between linguistic units and functional units in neurosciences. Finally, it makes the interesting suggestion that principles of morpheme-based organisation of the mental lexicon are compatible with a learning strategy requiring memorisation of full forms.}, KEYWORDS = {morphological structure, word processing, token/type frequency}, PAGES = {22-37}, URL = {http://mmm.lis.upatras.gr/index.php/mmm/issue/view/293/showToc}, PUBLISHER = {Università degli Studi di Bologna (Bologna, Italia)}, ISSN = {1826-7491}, CONFERENCE_NAME = {Morphology and Semantics-Ninth Mediterranean Morphology Meeting}, CONFERENCE_PLACE = {Dubrovnik (Croatia)}, CONFERENCE_DATE = {15-18/09/2013}, BOOKTITLE = {Morphology and Semantics}, EDITOR = {Audring, J. and Koutsoukos, N. and Masini, F. and Raffaelli, I.}, } @INPROCEEDINGS{MARZI_2015_INPROCEEDINGS_MFP_329352, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Lexical emergentism and the "frequency-by-regularity" interaction}, YEAR = {2015}, ABSTRACT = {In spite of considerable converging evidence of the role of inflectional paradigms in word acquisition and processing, little efforts have been put so far into providing detailed, algorithmic models of the interaction between lexical token frequency, paradigm frequency, paradigm regularity. We propose a neurocomputational account of this interaction, and discuss some theoretical implications of preliminary experimental results.}, KEYWORDS = {morphological strucutre, frequency distribution, temporal self-orgabnising maps}, PAGES = {37-41}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-84927156830\&origin=inward}, VOLUME = {1347}, PUBLISHER = {M. Jeusfeld c/o Redaktion Sun SITE, Informatik V, RWTH Aachen (Aachen, Germania)}, ISSN = {1613-0073}, CONFERENCE_NAME = {NetWordS Final Conference on Word Knowledge and Word Usage: Representations and Processes in the Mental Lexicon}, CONFERENCE_PLACE = {Pisa (Italy)}, CONFERENCE_DATE = {30-31/03 01/04 2015}, BOOKTITLE = {Word Knowledge and Word Usage 2015}, EDITOR = {Pirrelli, V. and Marzi, C. and Ferro, M.}, } @ARTICLE{CHERSI_2014_ARTICLE_CFPP_283372, AUTHOR = {Chersi, F. and Ferro, M. and Pezzulo, G. and Pirrelli, V.}, TITLE = {Topological Self-Organization and Prediction Learning Support Both Action and Lexical Chains in the Brain}, YEAR = {2014}, ABSTRACT = {A growing body of evidence in cognitive psychology and neuroscience suggests a deep interconnection between sensory-motor and language systems in the brain. Based on recent neurophysiological findings on the anatomo-functional organization of the fronto-parietal network, we present a computational model showing that language processing may have reused or co-developed organizing principles, functionality, and learning mechanisms typical of premotor circuit. The proposed model combines principles of Hebbian topological self-organization and prediction learning. Trained on sequences of either motor or linguistic units, the network develops independent neuronal chains, formed by dedicated nodes encoding only context-specific stimuli. Moreover, neurons responding to the same stimulus or class of stimuli tend to cluster together to form topologically connected areas similar to those observed in the brain cortex. Simulations support a unitary explanatory framework reconciling neurophysiological motor data with established behavioral evidence on lexical acquisition, access, and recall.}, KEYWORDS = {Motor chains, Lexical chains, Serial working memory, Computational modeling, Self-organizing maps, Somatotopic organization, Prediction}, PAGES = {476-491}, URL = {http://onlinelibrary.wiley.com/doi/10.1111/tops.12094/abstract?deniedAccessCustomisedMessage=\&userIsAuthenticated=false}, VOLUME = {6}, DOI = {10.1111/tops.12094}, PUBLISHER = {Cognitive Science Society, Inc (Hoboken, NJ, Stati Uniti d'America)}, ISSN = {1756-8757}, JOURNAL = {Topics in cognitive science (Print)}, } @ARTICLE{MARZI_2014_ARTICLE_MFK_288212, AUTHOR = {Marzi, C. and Ferro, M. and Keuleers, E.}, TITLE = {Perception of typicality in the lexicon: Wordlikeness, lexical density and morphonotactic constraints}, YEAR = {2014}, ABSTRACT = {The extent to which a symbolic time-series (a sequence of sounds or letters) is a typical word of a language, referred to as WORDLIKENESS, has been shown to have effects in speech perception and production, reading proficiency, lexical development and lexical access, short-term and long-term verbal memory. Two quantitative models have been suggested to account for these effects: serial phonotactic probabilities (the likelihood for a given symbolic sequence to appear in the lexicon) and lexical density (the extent to which other words can be obtained from a target word by changing, deleting or inserting one or more symbols in the target). The two measures are highly correlated and thus easy to be confounded in measuring their effects in lexical tasks. In this paper, we propose a computational model of lexical organisation, based on Self-Organising Maps with Hebbian connections defined over a temporal layer (TSOMs), providing a principled algorithmic account of effects of lexical acquisition, processing and access, to further investigate these issues. In particular, we show that (morpho-)phonotactic probabilities and lexical density, though correlated in lexical organisation, can be taken to focus on different aspects of speakers' word processing behaviour and thus provide independent cognitive contributions to our understanding of the principles of perception of typicality that govern lexical organisation.}, KEYWORDS = {wordlikeness, lexical access, word processing, frequency, memory}, PAGES = {171-191}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-84919701117\&origin=inward}, VOLUME = {40}, PUBLISHER = {Zavod za lingvistiku Filozofskog fakulteta (Zagreb, Croazia)}, ISSN = {0586-0296}, JOURNAL = {Suvremena lingvistika}, } @ARTICLE{MARZI_2014_ARTICLE_MFP_287289, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Morphological structure through lexical parsability}, YEAR = {2014}, ABSTRACT = {The emergence of morphological structure in lexical acquisition is analysed in the computational framework of Temporal Self-Organising Maps (TSOMs), to provide an explanatory basis for both psycholinguistic and linguistic accounts of lexical parsability. The investigation we propose is grounded on the hypothesis that perception of morphological structure (parsability) and frequency strongly correlate in the acquisition of inflectional paradigms. Analysis of experimental results of word acquisition obtained by artificially varying training conditions, allows us to understand developmental competition between fully-inflected word forms, and to investigate a hierarchy of frequency effects. The computational and theoretical implications of such a memory-based view of the relationship between frequency and perception, and its potential to account}, KEYWORDS = {inflectional paradigms, morphological structure, token/type frequency, word processing}, PAGES = {263-290}, URL = {http://www.rivisteweb.it/doi/10.1418/78410}, VOLUME = {XIII}, DOI = {10.1418/78410}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @INPROCEEDINGS{GAGGIOLI_2014_INPROCEEDINGS_GCSPTBCFCTDGTRR_283374, AUTHOR = {Gaggioli, A. and Cipresso, P. and Serino, S. and Pioggia, G. and Tartarisco, G. and Baldus, G. and Corda, D. and Ferro, M. and Carbonaro, N. and Tognetti, A. and De Rossi, D. and Giakoumis, D. and Tzovaras, D. and Riera, A. and Riva, G.}, TITLE = {A decision support system for real-time stress detection during virtual reality exposure}, YEAR = {2014}, ABSTRACT = {Virtual Reality (VR) is increasingly being used in combination with psycho-physiological measures to improve assessment of distress in mental health research and therapy. However, the analysis and interpretation of multiple physiological measures is time consuming and requires specific skills, which are not available to most clinicians. To address this issue, we designed and developed a Decision Support System (DSS) for automatic classification of stress levels during exposure to VR environments. The DSS integrates different biosensor data (ECG, breathing rate, EEG) and behavioral data (body gestures correlated with stress), following a training process in which self-rated and clinical-rated stress levels are used as ground truth. Detected stress events for each VR session are reported to the therapist as an aggregated value (ranging from 0 to 1) and graphically displayed on a diagram accessible by the therapist through a web-based interface.}, KEYWORDS = {Psychological Stress, Psychophysiology, Virtual Reality, Decision Support System, Biosensors}, PAGES = {114-120}, URL = {https://publications.cnr.it/doc/283374}, VOLUME = {196}, DOI = {10.3233/978-1-61499-375-9-114}, PUBLISHER = {IOS Press (Tokyo, Paesi Bassi)}, ISSN = {0926-9630}, CONFERENCE_NAME = {Medicine Meets Virtual Reality (MMVR21)}, BOOKTITLE = {Medicine Meets Virtual Reality}, EDITOR = {Westwood, J. D.}, } @INPROCEEDINGS{MARZI_2014_INPROCEEDINGS_MNF_295178, AUTHOR = {Marzi, C. and Nahli, O. and Ferro, M.}, TITLE = {Word Processing for Arabic Language: A reappraisal of morphology induction through adaptive memory self-organisation strategies}, YEAR = {2014}, ABSTRACT = {Modelling the mental lexicon focuses on processing and storage dynamics, since lexical organisation relies on the process of input recoding and adaptive strategies for long-term memory organisation. A fundamental issue in word processing is represented by the emergence of the morphological organisation level in the lexicon, based on paradigmatic relations between fully-stored word forms. Morphology induction can be defined as the task of identifying morphological formatives within morphologically complex word forms. In the computational framework we propose here (TSOMs), based on Self-Organising Maps with Hebbian connections defined over a temporal layer, the identification/perception of surface morphological relations involves the alignment of recoded representations of morphologically-related input words. Facing a non-concatenative morphology such as the Arabic inflectional system prompts a reappraisal of morphology induction through adaptive organisation strategies, which affect both lexical representations and long-term storage. We will show how a strongly adaptive self-organisation during training is conducive to emergent relations between stored word forms, and to high accuracy rates in generalising knowledge of stored words to unknown forms.}, KEYWORDS = {Non-concatenative morphological structure, lexical storage and access, SOMs, word recoding and processing, adaptive strategies, morphology}, PAGES = {241-247}, URL = {http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=7016626\&punumber%3D6996097}, DOI = {10.1109/CIST.2014.7016626}, PUBLISHER = {IEEE (New York, USA)}, ISBN = {978-1-4799-5979-2}, CONFERENCE_NAME = {Third IEEE International Colloquium in Information Science and Technology (CIST)}, CONFERENCE_PLACE = {Tetuan (Morocco)}, CONFERENCE_DATE = {20-22/10/2014}, BOOKTITLE = {IEEE Conference Publications-Catalog Number: CFP1467R-ART}, EDITOR = {El Mohajir, M. and Al Achhab, M. and Chahhou, M. and Mounir, A. and El Mohajir, B. and Pirrelli, V. and Zarghili, A. and Elfar, M.}, } @INPROCEEDINGS{PIRRELLI_2014_INPROCEEDINGS_PMF_290601, AUTHOR = {Pirrelli, V. and Marzi, C. and Ferro, M.}, TITLE = {Two-dimensional Wordlikeness Effects in Lexical Organisation}, YEAR = {2014}, ABSTRACT = {The main focus of research on wordlikeness has been on how serial processing strategies affect perception of similarity and, ultimately, the global network of associative relations among words in the mental lexicon. Comparatively little effort has been put so far, however, into an analysis of the reverse relationship: namely, how global organisation effects influence the speakers' perception of word similarity and of words' internal structure. In this paper, we explore the relationship between the two dimensions of wordlikeness (the "syntagmatic" and the "paradigmatic" one), to suggest that the same set of principles of memory organisation can account for both dimensions.}, KEYWORDS = {wordlikeness, lexical access, word processing, frequency, memory}, PAGES = {301-305}, URL = {http://clic.humnet.unipi.it/it/atti.html}, VOLUME = {1}, DOI = {10.12871/CLICIT2014158}, ISBN = {978-8-86741-472-7}, CONFERENCE_NAME = {First Italian Conference on Computational Linguistics CLiC-it 2014 \& Fourth International Workshop EVALITA 2014}, CONFERENCE_PLACE = {Pisa, Italy}, CONFERENCE_DATE = {9-11/12/2014}, BOOKTITLE = {The First Italian Conference on Computational Linguistics-Proceedings}, EDITOR = {Basili, R. and Lenci, A. and Magnini, B.}, } @ARTICLE{MARZI_2013_ARTICLE_MF_283382, AUTHOR = {Marzi, C. and Ferro, M.}, TITLE = {Adaptive strategies in lexical acquisition}, YEAR = {2013}, ABSTRACT = {The emergence of morphological patterns from lexical storage in language acquisition is conditioned by language-specific factors as well as extra-linguistic cognitive capacities. With particular reference to the acquisition of plural markers in German, in a memory-based perspective highlighting interesting theoretical implications for usage-based models, the paper analyses acquisitional strategies by focussing on emergent relations between stored word forms and on dynamic expectation/competition of incoming input. In particular, we outline an adaptive multifactorial account of morphological processing that includes both frequency and formal factors. Our investigation is supported by a computational model of morphology acquisition/processing based on self-organisation memories, where word representations are dynamically recoded as time-series.}, KEYWORDS = {German plurals, Morphological generalisation, Self-organising memory, Word processing}, PAGES = {307-328}, URL = {http://www.scopus.com/inward/record.url?eid=2-s2.0-84892521599\&partnerID=q2rCbXpz}, VOLUME = {XII}, DOI = {10.1418/75045}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @INPROCEEDINGS{MARZI_2013_INPROCEEDINGS_MFP_287555, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Lexical parsability and morphological structure}, YEAR = {2013}, ABSTRACT = {A classical tenet in the psycholinguistic literature on the mental lexicon is that a parsed affix presents high activation levels (and thus contributes to activation spreading to other words with the same affix), and that such levels are tightly correlated with the affix productivity. In a number of influential papers, it has been suggested that parsability criteria interact with frequency to define morphological productivity in the lexicon. For example, the frequency of a derivative (e.g. government) relative to its base (govern) is shown to be a good predictor for parsability/productivity. The higher the frequency ratio, the more likely the morphological structure to be perceived, and the associated affix to be used productively. The present contribution intends to offer a computational explanatory basis for this correlational evidence, and assess its applicability to the acquisition of complex inflectional paradigms. In those languages, like Italian and German, whose inflection is stem-based rather than word-based, there is often no single paradigmatic form which can act as a base by being properly contained in all other inflected variants. Yet, it seems intuitive to suggest that verbs that are inflected for one paradigm cell only (e.g. neighbouring), are learned earlier and more easily but exhibit lower levels of perceived inflectional structure than verbs with richer paradigms. This appears to be in good accord with experimental evidence of time latencies in lexical decision, which are shown to correlate negatively with token frequency, paradigm size and paradigm entropy. Our simulations, based on Temporal Self-Organizing Maps (TSOMs) allow us to establish an interesting connection between inflectional parsability, frequency-based paradigm structure, and acquisitional constraints on the interaction between the human processor and working memory. Self-organising topological models of the mental lexicon can mimic the spatial and temporal organization of memory structures supporting the processing of symbolic sequences [8-10], and can provide an interesting framework for testing integrative accounts of lexical processing/acquisition as the complex result of general-purpose operations on word stimuli (e.g. working memory, long-term storage, sensory-motor mapping, rehearsal, unit integration, unit analysis, executive control, time-series processing), in line with recent acquisitions on the neuro-functional architecture of the perisylvian language network in the left hemisphere of human brain. Simulations of the incremental acquisition of "mini-paradigms" (small islands of morphological contrast encompassing up to three different forms for the same verb support the hypothesis that perception of structure (parsability) and morphological productivity strongly correlate in the inflectional lexica of German and Italian. In particular, by monitoring longitudinal progress in storage and generalisation of differently distributed inflectional paradigms in the two languages, we show that: i) high-frequency forms are stored and accessed significantly earlier than low-frequency forms; ii) deeply entrenched but paradigmatically isolated forms tend to block usage of other forms in the same paradigm; iii) low-frequency evenly distributed (highly entropic) intra-paradigmatic forms are acquired later but are easily extended. Our investigation credits the proposed computational framework with psycholinguistic plausibility, and grounds parsability-based models of morphological productivity on a specific, explicit proposal of lexical architecture. This provides an explanatory basis for both psycholinguistic and linguistic accounts of morphological structure, and offers an intermediate framework for scientific inquiry bridging the gap between linguistic units and functional units in neurosciences. Finally, it makes the interesting suggestion that principles of morpheme-based organisation of the mental lexicon are compatible with a learning strategy requiring memorisation of full forms.}, KEYWORDS = {morphological structure, word paradigms, frequency, human processor}, PAGES = {33-34}, URL = {http://mmm9.ffzg.unizg.hr/wp-content/uploads/2012/10/MMM_PROGRAM4.pdf}, CONFERENCE_NAME = {9th Mediterannean Morphology Meeting on "Morphology and Semantics" (9th MMM)}, CONFERENCE_PLACE = {Dubrovnik, Croatia}, CONFERENCE_DATE = {15-18/09/2013}, BOOKTITLE = {Morphology and Semantics-Books of Abstracts}, } @INPROCEEDINGS{RUTA_2013_INPROCEEDINGS_RSTBCCBTFNMP_284784, AUTHOR = {Ruta, L. and Siracusano, R. and Tortorella, G. and Boncoddo, M. and Colombi, C. and Crifaci, G. and Billeci, L. and Tartarisco, G. and Ferro, M. and Narzisi, A. and Muratori, F. and Pioggia, G.}, TITLE = {The PRIMA-PIETRA Project: A Web-Based Platform for Early Autism Risk Assessment}, YEAR = {2013}, ABSTRACT = {It is well recognized that the best outcomes in autism spectrum disorders (ASD) are achieved through early diagnosis and early intervention. ASD symptoms may occur as early as 12-18 months and different instruments have been developed for early autism risk assessment under the age of 2 years. The Modified Checklist for Autism in Children (M-CHAT) is a developmental surveillance-screening instrument administered during 18- to 36-month well-child visits that was demonstrated to improve early identification of autism. Novel technologies can substantially contribute to improve early diagnosis in ASD, providing early screening risk assessment platforms, unobtrusive measurements of behaviors and physiological responses, as well as brain structure and connectivity, or other measurable stimulus-event experimental paradigms. The Prima Pietra Project based at the Pervasive Healthcare Center of the Institute of Clinical Physiology of the National Research Council of Italy (Consiglio Nazionale delle Ricerche, C.N.R.) and the AOU Polyclinic "G. Martino" in Messina developed and provided an early autism risk assessment web-based platform for pediatricians and physicians available on the internet.}, KEYWORDS = {early autism risk assessment}, URL = {https://imfar.confex.com/imfar/2013/webprogram/Paper14488.html}, CONFERENCE_NAME = {International Meeting for Autism Research 2013}, CONFERENCE_PLACE = {San Sebastian, Spain}, CONFERENCE_DATE = {2 May 2013}, BOOKTITLE = {International Meeting for Autism Research}, } @ARTICLE{MARZI_2012_ARTICLE_MFP_217399, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Word alignment and paradigm induction}, YEAR = {2012}, ABSTRACT = {The variety of morphological processes attested in inflectional system of average complexity calls for adaptive strategies of word alignment. Prefixation, suffixation, stem alternation and combinations thereof pose severe problems to unsupervised algorithms of morphology induction. The paper analyses morphological generalisation as a by-product of flexible memory self-organisation strategies for word recoding. Our model endorses the hypothesis that lexical forms are memorised as full units. At the same time, lexical units are paradigmatically organised. We show that the overall amount of redundant morphological structure emerging from paradigm-based self-organisation has a clear impact on generalisation. This supports the view that issues of word representation and issues of word processing are mutually implied in lexical acquisition.}, KEYWORDS = {Morphological Generalisation Morphological Paradigms Self-Organising Memory Word coding and Processing}, PAGES = {251-274}, URL = {http://www.rivisteweb.it/doi/10.1418/38789}, VOLUME = {XI}, DOI = {10.1418/38789}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @ARTICLE{TARTARISCO_2012_ARTICLE_TBCRAFGP_196454, AUTHOR = {Tartarisco, G. and Baldus, G. and Corda, D. and Raso, R. and Arnao, A. and Ferro, M. and Gaggioli, A. and Pioggia, G.}, TITLE = {Personal Health System architecture for stress monitoring and support to clinical decisions}, YEAR = {2012}, ABSTRACT = {Developments in computational techniques including clinical decision support systems, information processing, wireless communication and data mining hold new premises in Personal Health Systems. Pervasive Healthcare system architecture finds today an effective application and represents in perspective a real technological breakthrough promoting a paradigm shift from diagnosis and treatment of patients based on symptoms to diagnosis and treatment based on risk assessment. Such architectures must be able to collect and manage a large quantity of data supporting the physicians in their decision process through a continuous pervasive remote monitoring model aimed to enhance the understanding of the dynamic disease evolution and personal risk. In this work an automatic simple, compact, wireless, personalized and cost efficient pervasive architecture for the evaluation of the stress state of individual subjects suitable for prolonged stress monitoring during normal activity is described. A novel integrated processing approach based on an autoregressive model, artificial neural networks and fuzzy logic modeling allows stress conditions to be automatically identified with a mobile setting analysing features of the electrocardiographic signals and human motion. The performances of the reported architecture were assessed in terms of classification of stress conditions. © 2011 Elsevier B.V. All rights reserved.}, KEYWORDS = {Autonomic sympathovagal balance Autoregressive model Clinical decision support system Pervasive healthcare architecture Stress detection}, PAGES = {1296-1305}, URL = {http://www.sciencedirect.com/science/article/pii/S0140366411003720}, VOLUME = {35}, DOI = {10.1016/j.comcom.2011.11.015}, PUBLISHER = {IPC Science and Technology Press (Guildford, Regno Unito)}, ISSN = {0140-3664}, JOURNAL = {Computer communications}, } @ARTICLE{TERRANOVA_2012_ARTICLE_TFCRBSP_209663, AUTHOR = {Terranova, G. and Ferro, M. and Carpeggiani, C. and Recchia, V. and Braga, L. and Semelka, R. C. and Picano, E.}, TITLE = {Low Quality and Lack of Clarity of Current Informed Consent Forms in Cardiology: How to Improve Them}, YEAR = {2012}, ABSTRACT = {Guidelines on informed consent for clinical practice exhort physicians to use standard plain language to enhance patient comprehension and facilitate shared decision making. The aim of this study was to assess and improve quality and readability of current informed consent forms used in cardiology. We evaluated the currently used informed consent forms, previously written in Italian and English, of 7 common imaging examinations, according to the recommendations of scientific societies. For each text, we also developed a revised informed consent form according to reference standards, including Federal Plain Language guidelines. Regarding readability scores, we analyzed each text (standard and revised) with Flesch-Kincaid (F-K) grade level (higher numbers indicating harder-to-read text) and the Italian language-tailored Gulpease level (from 0 [difficult] to 100 [easy]). Overall quality and readability was poor for both the original English and Italian versions, and readability was improved with the revised form, with higher readability evidenced by changes in both F-K grade level (standard 10.2 ± 2.37% vs. revised 6.5 ± 0.41%; p < 0.001) for English and Gulpease (standard 45.7 ± 2% vs. revised 84.09 ± 2.98%; p < 0.0001) for Italian. In conclusion, current informed consent forms are complex, incomplete, and unreadable for the average patient. Substantial quality improvement and higher readability scores can be achieved with revised forms that explicitly discuss risks and are prepared following standard recommendations of plain writing.}, KEYWORDS = {bioethics imaging risk communication informed consent patient rights}, PAGES = {649-655}, URL = {http://www.sciencedirect.com/science/article/pii/S1936878X1200321X}, VOLUME = {5}, DOI = {10.1016/j.jcmg.2012.03.007}, PUBLISHER = {ELSEVIER SCIENCE INC (NEW YORK, USA, Stati Uniti d'America)}, ISSN = {1936-878X}, JOURNAL = {JACC-CARDIOVASCULAR IMAGING}, } @INCOLLECTION{IOANNDIS_2012_INCOLLECTION_ITDFVTP_218873, AUTHOR = {Ioanndis, D. and Tzovaras, D. and Dalle Mura, G. and Ferro, M. and Valenza, G. and Tognetti, A. and Pioggia, G.}, TITLE = {Gait and Anthropometric Profile Biometrics: A Step Forward}, YEAR = {2012}, ABSTRACT = {While a sharp debate is emerging about whether conventional biometric technology offers society any significant advantages over other forms of identification, and whether it constitutes a threat to privacy, technology is rapidly progressing. Politicians and the public are still discussing fingerprinting and iris scan, while scientists and engineers are already testing futuristic solutions. Second generation biometrics - which include multimodal biometrics, behavioural biometrics, dynamic face recognition, EEG and ECG biometrics, remote iris recognition, and other, still more astonishing, applications - is a reality which promises to overturn any current ethical standard about human identification. Robots which recognise their masters, CCTV which detects intentions, voice responders which analyse emotions: these are only a few applications in progress to be developed.}, KEYWORDS = {biometrics sensing seat}, PAGES = {105-127}, URL = {http://www.springer.com/social+sciences/applied+ethics/book/978-94-007-3891-1}, VOLUME = {11}, DOI = {10.1007/978-94-007-3892-8_5}, PUBLISHER = {Springer Verlag (Norwell MA, USA)}, ISBN = {978-94-007-3891-1}, BOOKTITLE = {Second Generation Biometrics: The Ethical, Legal and Social Context}, EDITOR = {Mordini, E. and Tzovaras, D.}, } @INCOLLECTION{PIRRELLI_2012_INCOLLECTION_PFC_136472, AUTHOR = {Pirrelli, V. and Ferro, M. and Calderone, B.}, TITLE = {Learning Paradigms in Time and Space: Computational Evidence from Romance Languages}, YEAR = {2012}, ABSTRACT = {In the linguistic literature, paradigms have enjoyed a hybrid status, half-way between entrenched patterns of lexical organization and processing structures enforcing global constraints on the output of traditional inflection rules. We describe here an original computational model of the mental lexicon where paradigmatic structures emerge through learning as the by-product of the endogenous dynamics of lexical memorization as competitive self-organization, based on the complementary principles of formal contrast (in space) and association biuniqueness (in time).}, KEYWORDS = {Computational model, Lexical memorization, Mental lexicon, Processing structures, Self-organizing maps}, PAGES = {135-157}, URL = {http://www.scopus.com/record/display.url?eid=2-s2.0-84921732430\&origin=inward}, DOI = {10.1093/acprof:oso/9780199589982.003.0008}, PUBLISHER = {Oxford University Press (Oxford, GBR)}, ISBN = {978-0-19-958998-2}, BOOKTITLE = {Morphological Autonomy: Perspectives for Romance Inflectional Morphology}, EDITOR = {Maiden, M. and Smith, J. C. and Goldbach, M. and Hinzelin, M.}, } @INPROCEEDINGS{MARZI_2012_INPROCEEDINGS_MFCP_287129, AUTHOR = {Marzi, C. and Ferro, M. and Caudai, C. and Pirrelli, V.}, TITLE = {Evaluating Hebbian Self-Organizing Memories for Lexical Representation and Access}, YEAR = {2012}, ABSTRACT = {The lexicon is the store of words in long-term memory. Any attempt at modelling lexical competence must take issues of string storage seriously. In the present contribution, we discuss a few desiderata that any biologically-inspired computational model of the mental lexicon has to meet, and detail a multi-task evaluation protocol for their assessment. The proposed protocol is applied to a novel computational architecture for lexical storage and acquisition, the "Topological Temporal Hebbian SOMs" (T2HSOMs), which are grids of topologically organised memory nodes with dedicated sensitivity to time-bound sequences of letters. These maps can provide a rigorous and testable conceptual framework within which to provide a comprehensive, multi-task protocol for testing the performance of Hebbian self-organising memories, and a comprehensive picture of the complex dynamics between lexical processing and the acquisition of morphological structure.}, KEYWORDS = {Mental Lexicon, Morphology Acquisition, Self-Organizing Maps}, PAGES = {886-893}, URL = {http://www.lrec-conf.org/proceedings/lrec2012/index.html}, ISBN = {978-2-9517408-7-7}, CONFERENCE_NAME = {8th International Conference on Language Resources and Evaluation}, CONFERENCE_PLACE = {Istanbul, Turkey}, CONFERENCE_DATE = {23-25/05/2012}, EDITOR = {Calzolari, N. and Choukri, K. and Declerck, T. and Uğur Doğan, M. and Maegaard, B. and Mariani, J. and Odijk, J. and Piperidis, S.}, } @INPROCEEDINGS{MARZI_2012_INPROCEEDINGS_MFP_219553, AUTHOR = {Marzi, C. and Ferro, M. and Pirrelli, V.}, TITLE = {Prediction and Generalisation in Word Processing and Storage}, YEAR = {2012}, ABSTRACT = {Word storage and processing have traditionally been modelled according to different computational paradigms, in line with the classical corner-stone of "dual-route" models of word structure assuming a sharp dissociation between memory and computation (Clahsen 1999, Di Sciullo \& Williams 1987, Pinker \& Prince 1988, Parasada \& Pinker 1993). Even the most radical alternative to dual-route thinking, connectionist one-route models, challenged the lexicon-grammar dualism only by providing a neurally-inspired mirror image of classical base-to-inflection rules, while largely neglecting issues of lexical storage (Rumelhart \& McClelland 1986, McClelland \& Patterson 2002, Seidenberg \& McClelland 1989). Recent psycho- and neuro-linguistic evidence, however, supports a less deterministic and modular view of the interaction between stored word knowledge and on-line processing [Baayen et al. 1997, Hay 2001, Maratsos 2000, Stemberger \& Middleton 2003, Tabak et al. 2005, Ford et al. 2003, Post et al. 2008). The view entails simultaneous activation of distributed patterns of cortical connectivity encoding redundant distributional regularities in language data. Furthermore, recent developments in morphological theorising question the primacy of grammar rules over lexical storage, arguing that word regularities emerge from independent principles of lexical organisation, whereby lexical units and constructions are redundantly stored and mutually related through entailment relations (Matthews 1991, Corbett \& Fraser 1993, Pirrelli 2000, Burzio 2004, Booij 2010). We endorse here such a non modular view on Morphology to investigate two basic behavioural aspects of human word processing: morphological prediction and generalisation. The investigation is based on a computer model of morphology acquisition supporting the hypothesis that they both derive from a common pool of principles of lexical organisation.}, KEYWORDS = {Morphological generalisation, Word processing, Self-organising memory}, PAGES = {114-131}, URL = {http://mmm.lingue.unibo.it/}, CONFERENCE_NAME = {Eighth Mediterranean Morphology Meeting on "Morphology and the architecture of the grammar" (MMM8)}, CONFERENCE_PLACE = {Cagliari, Italy}, CONFERENCE_DATE = {14-17 September 2011}, EDITOR = {Ralli, A. and Booij, G. and Scalise, S. and Karasimos, A.}, } @INPROCEEDINGS{PIOGGIA_2012_INPROCEEDINGS_PBNFATFSGDTM_284777, AUTHOR = {Pioggia, G. and Billeci, L. and Narzisi, A. and Farruggio, V. and Arnao, A. and Tartarisco, G. and Ferro, M. and Siracusano, R. and Germanò, E. and Deodato, M. and Tortorella, G. and Muratori, F.}, TITLE = {PRIMA PIETRA: Research, Integration, Enhancement, Assistance and Education Program for Autism Services and Rehabilitation Technologies}, YEAR = {2012}, ABSTRACT = {It is commonly recognized that autism spectrum disorder (ASD) symptoms are as early as 12 months of age and that the best outcomes are often achieved through early diagnosis and early intervention. However, there are many challenges to delivering health care to parents with a child with ASD. Difficulties to service delivery and utilization are more intensified for families living in suburban or remote areas, often resulting in limited access to preventative mental health services in general and parenting ASD interventions in particular. As Vismara an Rogers suggested (Vismara, 2010), the use of technology could support long-distance clinical health care. PRIMA PIETRA Italian project is focused on early diagnosis and intervention providing Early Start Denver Model (Dawson et al., 2009) using tele-rehabilitation. PRIMA PIETRA is a collaborative project supported by the Minister of Health of the Sicilian Region, in collaboration with Basilicata and Tuscany Regions.}, KEYWORDS = {autism spectrum disorder, pervasive healthcare, early diagnosis}, PAGES = {4}, URL = {https://imfar.confex.com/imfar/2012/webprogram/Paper10070.html}, CONFERENCE_NAME = {International Meeting for Autism Research}, CONFERENCE_PLACE = {Toronto, Canada}, CONFERENCE_DATE = {18 May 2012}, } @TECHREPORT{TARTARISCO_2012_TECHREPORT_TBCFP_221736, AUTHOR = {Tartarisco, G. and Baldus, G. and Corda, D. and Ferro, M. and Pioggia, G.}, TITLE = {Decision Support Processing Architecture}, YEAR = {2012}, ABSTRACT = {This report presents the design and im-plementation of the INTERSTRESS Deci-sion Support System (DSS). The goal of the DSS is to assess the psychological state of each patient by analyzing the previously acquired knowledge, such as patient's physiological and behavioural profile, and current sensory data. Starting from such information, the DSS then infers physiological and behavioural markers of stress.}, KEYWORDS = {decision support system multimodal analysis artificial neural networks Bayesian models machine learning}, URL = {https://publications.cnr.it/doc/221736}, } @TECHREPORT{VISINTAINER_2012_TECHREPORT_VMCKCPTF_221683, AUTHOR = {Visintainer, F. and Muro, M. and Carlino, A. and Kalogirou, K. and Contreras, J. and Pioggia, G. and Tartarisco, G. and Ferro, M.}, TITLE = {Two vehicle demonstrators for elderly drivers support}, YEAR = {2012}, KEYWORDS = {elderly support biometrics sensing seat}, URL = {https://publications.cnr.it/doc/221683}, } @ARTICLE{CHERSI_2011_ARTICLE_CFPP_205122, AUTHOR = {Chersi, F. and Ferro, M. and Pezzulo, G. and Pirrelli, V.}, TITLE = {Time, Language and Action-A Unified Long-Term Memory Model for Sensory-Motor Chains and Word Schemata}, YEAR = {2011}, ABSTRACT = {Action and language are known to be organized as closely-related brain subsystems. An Italian CNR project implemented a computational neural model where the ability to form chains of goal-directed actions and chains of linguistic units relies on a unified memory architecture obeying the same organizing principles.}, PAGES = {27-28}, URL = {http://ercim-news.ercim.eu/images/stories/EN84/EN84-web.pdf}, VOLUME = {84}, PUBLISHER = {ERCIM (Le Chesnay)}, ISSN = {0926-4981}, JOURNAL = {ERCIM news}, } @ARTICLE{FERRO_2011_ARTICLE_FMP_205180, AUTHOR = {Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {A Self-Organizing Model of Word Storage and Processing: Implications for Morphology Learning}, YEAR = {2011}, ABSTRACT = {In line with the classical cornerstone of "dual-route" models of word structure, assuming a sharp dissociation between memory and computation, word storage and processing have traditionally been modelled according to different computational paradigms. Even the most popular alternative to dual-route thinking - connectionist one-route models - challenged the lexicon-grammar dualism only by providing a neurally-inspired mirror image of classical base-to-inflection rules, while largely neglecting issues of lexical storage. Recent psycho- and neuro-linguistic evidence, however, supports a less deterministic and modular view of the interaction between stored word knowledge and on-line processing. We endorse here such a non modular view on morphology to offer a computer model supporting the hypothesis that they are both derivative of a common pool of principles for memory self-organization.}, KEYWORDS = {Lexical Processing, Self Organizing Maps, Morphological Structure, Serial Memory}, PAGES = {209-226}, URL = {http://www.rivisteweb.it/doi/10.1418/35840}, VOLUME = {2}, DOI = {10.1418/35840}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @ARTICLE{VALENZA_2011_ARTICLE_VPAFSD_329628, AUTHOR = {Valenza, G. and Pioggia, G. and Armato, A. and Ferro, M. and Scilingo, E. P. and De Rossi, D.}, TITLE = {A neuron-astrocyte transistor-like model for neuromorphic dressed neurons}, YEAR = {2011}, ABSTRACT = {Experimental evidences on the role of the synaptic glia as an active partner together with the bold synapse in neuronal signaling and dynamics of neural tissue strongly suggest to investigate on a more realistic neuron-glia model for better understanding human brain processing. Among the glial cells, the astrocytes play a crucial role in the tripartite synapsis, i.e. the dressed neuron. A well-known two-way astrocyte-neuron interaction can be found in the literature, completely revising the purely supportive role for the glia. The aim of this study is to provide a computationally efficient model for neuron-glia interaction. The neuron-glia interactions were simulated by implementing the Li-Rinzel model for an astrocyte and the Izhikevich model for a neuron. Assuming the dressed neuron dynamics similar to the nonlinear input-output characteristics of a bipolar junction transistor, we derived our computationally efficient model. This model may represent the fundamental computational unit for the development of real-time artificial neuron-glia networks opening new perspectives in pattern recognition systems and in brain neurophysiology.}, KEYWORDS = {Bio-computational architectures for signal processing, Neuron, Astrocyte, Synapse, Neuron-astrocyte interaction model}, PAGES = {679-685}, URL = {http://www.sciencedirect.com/science/article/pii/S0893608011000979}, VOLUME = {24}, DOI = {10.1016/j.neunet.2011.03.013}, PUBLISHER = {Pergamon (New York, Stati Uniti d'America)}, ISSN = {0893-6080}, JOURNAL = {Neural networks}, } @INPROCEEDINGS{FERRO_2011_INPROCEEDINGS_FMP_205490, AUTHOR = {Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {T2HSOM: Understanding the Lexicon by Simulating Memory Processes for Serial Order}, YEAR = {2011}, ABSTRACT = {Over the last several years, both theoretical and empirical approaches to lexical knowledge and encoding have prompted a radical reappraisal of the traditional dichotomy between lexicon and grammar. The lexicon is not simply a large waste basket of exceptions and sub-regularities, but a dynamic, possibly redundant repository of linguistic knowledge whose principles of relational organization are the driving force of productive generalizations. In this paper, we overview a few models of dynamic lexical organization based on neural network architectures that are purported to meet this challenging view. In particular, we illustrate a novel family of Kohonen self-organizing maps (T2HSOMs) that have the potential of simulating competitive storage of symbolic time series while exhibiting interesting properties of morphological organization and generalization. The model, tested on training samples of as morphologically diverse languages as Italian, German and Arabic, shows sensitivity to manifold types of morphological structure and can be used to bootstrap morphological knowledge in an unsupervised way.}, KEYWORDS = {Mental Lexicon, Self-organizing Maps, Morphology}, PAGES = {32-41}, URL = {http://alpage.inria.fr/~sagot/woler2011/WoLeR2011/Program_%26_Proceedings.html}, CONFERENCE_NAME = {First International Workshop on Lexical Resources}, CONFERENCE_PLACE = {Ljubljana Slovenia}, CONFERENCE_DATE = {1-5 Agosto 2011}, BOOKTITLE = {First International Workshop on Lexical Resources}, EDITOR = {Sagot, B.}, } @INPROCEEDINGS{TERRANOVA_2011_INPROCEEDINGS_TFCRDGRP_205909, AUTHOR = {Terranova, G. and Ferro, M. and Carpeggiani, C. and Recchia, V. and Dodaro, A. and Gioffrè, D. and Richard, S. and Picano, E.}, TITLE = {Unreadability of current informed consent forms in cardiology-and how to improve it}, YEAR = {2011}, ABSTRACT = {Guidelines on informed consent for clinical practice and research trials recommend the use of standard plain language to enhance patient comprehension and to facilitate shared decision-making. Aim: To assess readability of our current informed consent forms used in cardiology.}, KEYWORDS = {public health, health policy, informed consent, readability}, PAGES = {69-70}, URL = {http://spo.escardio.org/abstract-book/presentation.aspx?id=97162}, VOLUME = {32}, CONFERENCE_NAME = {European Society of Cardiology}, CONFERENCE_PLACE = {Paris (France)}, CONFERENCE_DATE = {Agosto 2011}, } @ARTICLE{FERRO_2010_ARTICLE_FOPP_64549, AUTHOR = {Ferro, M. and Ognibene, D. and Pezzulo, G. and Pirrelli, V.}, TITLE = {Reading as active sensing: a computational model of gaze planning in word recognition}, YEAR = {2010}, ABSTRACT = {We offer a computational model of gaze planning during reading that consists of two main components: a lexical representation network, acquiring lexical representations from input texts (a subset of the Italian CHILDES database), and a gaze planner, designed to recognize written words by mapping strings of characters onto lexical representations. The model implements an active sensing strategy that selects which characters of the input string are to be fixated, depending on the predictions dynamically made by the lexical representation network. We analyze the developmental trajectory of the system in performing the word recognition task as a function of both increasing lexical competence, and correspondingly increasing lexical prediction ability. We conclude by discussing how our approach can be scaled up in the context of an active sensing strategy applied to a robotic setting.}, KEYWORDS = {Reading, Language Learning, Mental Lexicon}, PAGES = {1-16}, URL = {https://publications.cnr.it/doc/64549}, VOLUME = {4}, PUBLISHER = {Frontiers Research Foundation (Lausanne, Svizzera)}, ISSN = {1662-5218}, JOURNAL = {Frontiers in neurorobotics}, } @ARTICLE{FERRO_2010_ARTICLE_FPP_64553, AUTHOR = {Ferro, M. and Pezzulo, G. and Pirrelli, V.}, TITLE = {Morphology, Memory and the Mental Lexicon}, YEAR = {2010}, ABSTRACT = {Recent experimental evidence on morphological learning and processing has prompted a less deterministic and modular view of the interaction between stored word knowledge and on-line processing. Storing a word in the mental lexicon does not simply entail keeping a faithful memory image of that word in the most compact way. It also requires encoding and manipulating such image through topological structures that are optimally adapted to word production and comprehension. Temporal Self-Organizing Maps (THSOMs) are a novel model of artificial neural network that keeps time serial information through predictive activation chains of receptors encoding both spatial and temporal information of input stimuli. The impact of this model on issues of lexical organization and morphological processing is investigated in detail through a series of simulations shedding light on the dynamics between short-term memory (activation), long-term memory (learning) and morphological organization of stored word forms (topology).}, KEYWORDS = {Morphology, Word Processing, Word Learning, Mental Lexicon}, PAGES = {203-242}, URL = {https://publications.cnr.it/doc/64553}, VOLUME = {2}, PUBLISHER = {Il Mulino, Bologna (Italia)}, ISSN = {1720-9331}, JOURNAL = {Lingue e linguaggio}, } @BOOK{FERRO_2010_BOOK_F_283384, AUTHOR = {Ferro, M.}, TITLE = {High Efficiency Real-Time Sensor and Actuator Control and Data Processing: A Framework Solution for Control Systems in Biomimetic Autonomous Robots}, YEAR = {2010}, URL = {https://publications.cnr.it/doc/283384}, PUBLISHER = {VDM Verlag Dr. Müller (Saarbrücken, DEU)}, ISBN = {978-3-639-25356-6}, EDITOR = {Ferro, M.}, } @MISC{FERRO_2010_MISC_FMP_157477, AUTHOR = {Ferro, M. and Marzi, C. and Pirrelli, V.}, TITLE = {Word self-organization in time and space? Algorithms and evaluation}, YEAR = {2010}, ABSTRACT = {ABSTRACT: Words are time-bound signals and are amenable to temporal processing. The human brain has an innate ability to encode serial events into spatial patterns of neural activity (David Beiser \& James Houk, 1998). Temporal Hebbian SOMs (THSOMs) allow us to take the two assumptions seriously. They provide a novel computational framework accounting for many paradigm-based generalizations in a natural and insightful way. This claim is validated on inflectional data from German, English and Italian.}, KEYWORDS = {Morphology, Word Processing and Learning, Mental Lexicon, L1, SOMs}, URL = {https://publications.cnr.it/doc/157477}, } @INCOLLECTION{FERRO_2009_INCOLLECTION_FP_283387, AUTHOR = {Ferro, M. and Pioggia, G.}, TITLE = {A biologically-based framework for distributed sensory fusion and data processing}, YEAR = {2009}, PAGES = {337-364}, URL = {https://publications.cnr.it/doc/283387}, DOI = {10.5772/6586}, ISBN = {978-3-902613-52-3}, BOOKTITLE = {Sensor and Data Fusion}, EDITOR = {Milisavljevic, N.}, }