{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/a-cognitive-neural-architecture-able-to-learn","title":"A cognitive neural architecture able to learn and communicate through natural language","arxiv_id":"1506.03229","date":"2015-06-10","proceeding":null,"authors":["Bruno Golosio","Angelo Cangelosi","Olesya Gamotina","Giovanni Luca Masala"],"abstract":"Communicative interactions involve a kind of procedural knowledge that is\nused by the human brain for processing verbal and nonverbal inputs and for\nlanguage production. Although considerable work has been done on modeling human\nlanguage abilities, it has been difficult to bring them together to a\ncomprehensive tabula rasa system compatible with current knowledge of how\nverbal information is processed in the brain. This work presents a cognitive\nsystem, entirely based on a large-scale neural architecture, which was\ndeveloped to shed light on the procedural knowledge involved in language\nelaboration. The main component of this system is the central executive, which\nis a supervising system that coordinates the other components of the working\nmemory. In our model, the central executive is a neural network that takes as\ninput the neural activation states of the short-term memory and yields as\noutput mental actions, which control the flow of information among the working\nmemory components through neural gating mechanisms. The proposed system is\ncapable of learning to communicate through natural language starting from\ntabula rasa, without any a priori knowledge of the structure of phrases,\nmeaning of words, role of the different classes of words, only by interacting\nwith a human through a text-based interface, using an open-ended incremental\nlearning process. It is able to learn nouns, verbs, adjectives, pronouns and\nother word classes, and to use them in expressive language. The model was\nvalidated on a corpus of 1587 input sentences, based on literature on early\nlanguage assessment, at the level of about 4-years old child, and produced 521\noutput sentences, expressing a broad range of language processing\nfunctionalities.","url_abs":"http://arxiv.org/abs/1506.03229v3","url_pdf":"http://arxiv.org/pdf/1506.03229v3.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"a-cognitive-neural-architecture-able-to-learn","repo_url":"https://github.com/golosio/annabell","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"incremental-learning","task_name":"Incremental Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}