{ "access": { "embargo": { "active": false, "reason": null }, "files": "public", "record": "public", "status": "open" }, "created": "2020-10-22T11:43:49.815132+00:00", "custom_fields": { "journal:journal": { "issue": "1, 2016", "pages": "27\u201351", "title": "Adaptive Behavior", "volume": "24" } }, "deletion_status": { "is_deleted": false, "status": "P" }, "files": { "count": 1, "enabled": true, "entries": { "Learning programs is better than learning.pdf": { "checksum": "md5:8b065c35f9ddafa60a16464441b67922", "ext": "pdf", "id": "dacf003d-d67e-4230-b515-12a60c718dbc", "key": "Learning programs is better than learning.pdf", "metadata": null, "mimetype": "application/pdf", "size": 5453249 } }, "order": [], "total_bytes": 5453249 }, "id": "4117980", "is_draft": false, "is_published": true, "links": { "access": "https://zenodo.org/api/records/4117980/access", "access_links": "https://zenodo.org/api/records/4117980/access/links", "access_request": "https://zenodo.org/api/records/4117980/access/request", "access_users": "https://zenodo.org/api/records/4117980/access/users", "archive": "https://zenodo.org/api/records/4117980/files-archive", "archive_media": "https://zenodo.org/api/records/4117980/media-files-archive", "communities": "https://zenodo.org/api/records/4117980/communities", "communities-suggestions": "https://zenodo.org/api/records/4117980/communities-suggestions", "doi": "https://doi.org/10.1177/1059712315609412", "draft": "https://zenodo.org/api/records/4117980/draft", "files": "https://zenodo.org/api/records/4117980/files", "latest": "https://zenodo.org/api/records/4117980/versions/latest", "latest_html": "https://zenodo.org/records/4117980/latest", "media_files": "https://zenodo.org/api/records/4117980/media-files", "parent": "https://zenodo.org/api/records/4117979", "parent_doi": "https://zenodo.org/doi/", "parent_html": "https://zenodo.org/records/4117979", "requests": "https://zenodo.org/api/records/4117980/requests", "reserve_doi": "https://zenodo.org/api/records/4117980/draft/pids/doi", "self": "https://zenodo.org/api/records/4117980", "self_doi": "https://zenodo.org/doi/10.1177/1059712315609412", "self_html": "https://zenodo.org/records/4117980", "self_iiif_manifest": "https://zenodo.org/api/iiif/record:4117980/manifest", "self_iiif_sequence": "https://zenodo.org/api/iiif/record:4117980/sequence/default", "versions": "https://zenodo.org/api/records/4117980/versions" }, "media_files": { "count": 0, "enabled": false, "entries": {}, "order": [], "total_bytes": 0 }, "metadata": { "creators": [ { "affiliations": [ { "name": "Institute of Cognitive Sciences and Technologies, National Research Council of Italy, Italy" } ], "person_or_org": { "family_name": "Francesco Donnarumma", "identifiers": [ { "identifier": "0000-0003-4248-5360", "scheme": "orcid" } ], "name": "Francesco Donnarumma", "type": "personal" } }, { "affiliations": [ { "name": "Dipartimento di Ingegneria Elettrica e Tecnologie dell'Informazione, Universit\u00e0 degli Studi di Napoli Federico, Napoli" } ], "person_or_org": { "family_name": "Roberto Prevete", "identifiers": [ { "identifier": "0000-0002-3804-1719", "scheme": "orcid" } ], "name": "Roberto Prevete", "type": "personal" } }, { "affiliations": [ { "name": "KTH Royal Institute of Technology" } ], "person_or_org": { "family_name": "Andrea de Giorgio", "identifiers": [ { "identifier": "0000-0001-6064-5634", "scheme": "orcid" } ], "name": "Andrea de Giorgio", "type": "personal" } }, { "affiliations": [ { "name": "Institut Neuroscience Cognition, Universit\u00e9 Paris Descartes, France" } ], "person_or_org": { "family_name": "Guglielmo Montone", "name": "Guglielmo Montone", "type": "personal" } }, { "affiliations": [ { "name": "Institute of Cognitive Sciences and Technologies, National Research Council of Italy, Italy" } ], "person_or_org": { "family_name": "Giovanni Pezzulo", "identifiers": [ { "identifier": "0000-0001-6813-8282", "scheme": "orcid" } ], "name": "Giovanni Pezzulo", "type": "personal" } } ], "description": "
Distributed and hierarchical models of control are nowadays popular in computational modeling and robotics. In the artificial neural network literature, complex behaviors can be produced by composing elementary building blocks or motor primitives, possibly organized in a layered structure. However, it is still unknown how the brain learns and encodes multiple motor primitives, and how it rapidly reassembles, sequences and switches them by exerting cognitive control. In this paper we advance a novel proposal, a hierarchical programmable neural network architecture, based on the notion of programmability and an interpreter-programmer computational scheme. In this approach, complex (and novel) behaviors can be acquired by embedding multiple modules (motor primitives) in a single, multi-purpose neural network. This is supported by recent theories of brain functioning in which skilled behaviors can be generated by combining functional different primitives embedded in “reusable” areas of “recycled” neurons. Such neuronal substrate supports flexible cognitive control, too. Modules are seen as interpreters of behaviors having controlling input parameters, or programs that encode structures of networks to be interpreted. Flexible cognitive control can be exerted by a programmer module feeding the interpreters with appropriate input parameters, without modifying connectivity. Our results in a multiple T -maze robotic scenario show how this computational framework provides a robust, scalable and flexible scheme that can be iterated at different hierarchical layers permitting to learn, encode and control multiple qualitatively different behaviors.
", "languages": [ { "id": "eng", "title": { "en": "English" } } ], "publication_date": "2015-10-29", "publisher": "Zenodo", "resource_type": { "id": "publication-article", "title": { "de": "Zeitschriftenartikel", "en": "Journal article" } }, "rights": [ { "description": { "en": "The Creative Commons Attribution license allows re-distribution and re-use of a licensed work on the condition that the creator is appropriately credited." }, "icon": "cc-by-icon", "id": "cc-by-4.0", "props": { "scheme": "spdx", "url": "https://creativecommons.org/licenses/by/4.0/legalcode" }, "title": { "en": "Creative Commons Attribution 4.0 International" } } ], "subjects": [ { "subject": "Artificial neural networks" }, { "subject": "Programmable neural networks" }, { "subject": "Machine learning" }, { "subject": "Robotics" } ], "title": "Learning programs is better than learning dynamics: A programmable neural network hierarchical architecture in a multi-task scenario" }, "parent": { "access": { "owned_by": { "user": 146569 } }, "communities": { "default": "455b2f5a-1010-4bab-8961-35815c92615e", "entries": [ { "access": { "member_policy": "open", "members_visibility": "public", "record_policy": "open", "review_policy": "open", "visibility": "public" }, "children": { "allow": false }, "created": "2020-10-24T10:53:30.056758+00:00", "custom_fields": {}, "deletion_status": { "is_deleted": false, "status": "P" }, "id": "455b2f5a-1010-4bab-8961-35815c92615e", "links": {}, "metadata": { "curation_policy": "In order for content to be approved, at least one author must have an affiliation with a university or company based in Stockholm. Content must be either in English or Swedish language.
\r\n\r\nStockholm AI reserves its right to not publish or remove any content that violates the curation policy.
\r\n", "page": "Stockholm AI is a grassroots movement where Artificial Intelligence (AI) enthusiasts, professionals, researchers and students come together with the purpose of establishing Stockholm as a global player in the field of AI.
\r\n\r\nThe Scientific Repository is curated as a repost journal on Artificial Intelligence research and applications, in any fields. It means that any AI-related open access content published in scientific journals or conference proceedings can be reposted here and shared with the Stockholm AI community. But not only those: Code, posters, drafts, reports, presentations, videos, datasets, books, theses & more are always more than welcome!
\r\n\r\n\r\n\r\n