MULTIPLEX LEXICAL NETWORKS AS MODELS OF LANGUAGE REPRESENTATION IN THE HUMAN MIND скачать в хорошем качестве

MULTIPLEX LEXICAL NETWORKS AS MODELS OF LANGUAGE REPRESENTATION IN THE HUMAN MIND

-Massimo Stella, PhD in Complex Systems Simulation, MSc in Physics, Complex Data Scientist, Founder, Italy https://www.researchgate.net/profile/... This lecture revolves around multiplex lexical networks as quantitative tools for investigating how multiple types of conceptual associations can simultaneously influence language acquisition and use. The main object of investigation is the mental lexicon, a cognitive system apt at language learning, storage, processing and production. This system is networked and multiplex: Years of empirical research have shown how the layouts of semantic and phonological associations between concepts influence word processing in interactive ways, potentially in synergy also with other aspects of language like orthography, syntax and emotional meaning. In agreement with past models of language representation in the human mind, multiplex networks provide quantitative tools for modelling how multiple types of conceptual associations across thousands of words/concepts influence our cognitive reflection of language and its dynamics in phenomena like early word acquisition, phonological priming or word production in people with cognitive impairments. This lecture will introduce basic definitions from multiplex networks, highlighting how multiplexity alters metrics and concepts originally developed for single-layer networks. Particular attention will be devoted to drawing a parallel between cognitive interpretation and quantitative formalism, including applications in machine learning. Further, a few case studies from the recent literature will be reviewed and discussed, highlighting also recently available large-scale datasets of interest, like specific types of networks, language norms and labelled datasets. The online workshop for young scientists “Network analysis in human cognition and language processing” Воркшоп Института психологии РАН для молодых ученых "Моделирование и анализ сетей в когнитивной системе человека и языке". 24 - 25 октября 2020 года