A neural network architecture for learning word–referent associations in multiple contexts

Abstract

This article proposes a biologically inspired neurocomputational architecture which learns associations between words and referents in different contexts, considering evidence collected from the literature of Psycholinguistics and Neurolinguistics. The multi-layered architecture takes as input raw images of objects (referents) and streams of word's phonemes (labels), builds an adequate representation, recognizes the current context, and associates label with referents incrementally, by employing a Self-Organizing Map which creates new association nodes (prototypes) as required, adjusts the existing prototypes to better represent the input stimuli and removes prototypes that become obsolete/unused. The model takes into account the current context to retrieve the correct meaning of words with multiple meanings. Simulations show that the model can reach up to 78% of word-referent association accuracy in ambiguous situations and approximates well the learning rates of humans as reported by three different authors in five Cross-Situational Word Learning experiments, also displaying similar learning patterns in the different learning conditions.

Word-Object Learning via Visual Exploration in Space (WOLVES): A neural process model of cross-situational word learning.

“…Furthermore, like HT models of CSWL, the majority of AL models have been developed in the context of, and applied to, single empirical studies (see Yu & Ballard, 2007; Yurovsky et al, 2014 in Table 1) or limited sets of data such as small utterance corpuses rather than the results of empirical studies (see Fazly et al, 2010; Yu & Smith, 2011; Nematzadeh et al, 2012 in Table 1). A few AL models have captured data from multiple studies (Bassani & Araujo, 2019; Kachergis et al, 2012, 2013, 2017; Räsänen & Rasilo, 2015), suggesting that they are better able to generalize across specific CSWL paradigms. Yet, although promising, no AL model has been applied to the full range of CSWL studies from infants to adults, and thus, no AL account has explained changes in CSWL over development.…”

Word-Object Learning via Visual Exploration in Space (WOLVES): A Neural Process Account of Cross-Situational Word Learning

“…Furthermore, like HT models of CSWL, the majority of AL models have been developed in the context of, and applied to, single empirical studies (see Yu & Ballard, 2007;Yurovsky, Fricker, Yu & Smith, 2014 in Table 1) or limited sets of data such as small utterance corpuses rather than the results of empirical studies (see Fazly et al;Yu & Smith, 2011;Nematzadeh, Fazly & Stevenson, 2012 in Table 1). A few AL models have captured data from multiple studies (Bassani & Araujo, 2019;Kachergis et al, 2012;Kachergis, Yu, & Shiffrin, 2013Rasanen & Rasilo, 2015), suggesting that they are better able to generalize across specific CSWL paradigms. Yet, while promising, no AL model has been applied to the full range of CSWL studies from infants to adults, and thus no AL account has explained changes in CSWL over development.…”

Artificial Neural Networks with Turkey Interconnected System for Day Ahead Market Transmission Lines Estimated Losses

“…Hetereoasosyatif YSA yapısına sahip olan ÇKA birden fazla katmana sahiptir. ÇKA, doğrusal olmayan problemleri çözebilmeleri nedeniyle günümüzde geniş kullanım alanları bulan en popüler yapay sinir ağıdır (Bassani & Araujo, 2019;Ban & Chang, 2015;Yıldız vd., 2019). Ayrıca kullandığı öğrenme algoritması nedeniyle geri yayılım ağı olarak da anılmaktadır (Sarıgül vd., 2019).…”