{"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-modular-architecture-for-transparent","title":"A modular architecture for transparent computation in Recurrent Neural Networks","arxiv_id":"1609.01926","date":"2016-09-07","proceeding":null,"authors":["Giovanni Sirio Carmantini","Peter beim Graben","Mathieu Desroches","Serafim Rodrigues"],"abstract":"Computation is classically studied in terms of automata, formal languages and\nalgorithms; yet, the relation between neural dynamics and symbolic\nrepresentations and operations is still unclear in traditional eliminative\nconnectionism. Therefore, we suggest a unique perspective on this central\nissue, to which we would like to refer as to transparent connectionism, by\nproposing accounts of how symbolic computation can be implemented in neural\nsubstrates. In this study we first introduce a new model of dynamics on a\nsymbolic space, the versatile shift, showing that it supports the real-time\nsimulation of a range of automata. We then show that the Goedelization of\nversatile shifts defines nonlinear dynamical automata, dynamical systems\nevolving on a vectorial space. Finally, we present a mapping between nonlinear\ndynamical automata and recurrent artificial neural networks. The mapping\ndefines an architecture characterized by its granular modularity, where data,\nsymbolic operations and their control are not only distinguishable in\nactivation space, but also spatially localizable in the network itself, while\nmaintaining a distributed encoding of symbolic representations. The resulting\nnetworks simulate automata in real-time and are programmed directly, in absence\nof network training. To discuss the unique characteristics of the architecture\nand their consequences, we present two examples: i) the design of a Central\nPattern Generator from a finite-state locomotive controller, and ii) the\ncreation of a network simulating a system of interactive automata that supports\nthe parsing of garden-path sentences as investigated in psycholinguistics\nexperiments.","url_abs":"http://arxiv.org/abs/1609.01926v1","url_pdf":"http://arxiv.org/pdf/1609.01926v1.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-modular-architecture-for-transparent","repo_url":"https://github.com/TuringMachinegun/Turing_Neural_Networks","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}