{"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/logical-learning-through-a-hybrid-neural","title":"Logical Learning Through a Hybrid Neural Network with Auxiliary Inputs","arxiv_id":"1705.08200","date":"2017-05-23","proceeding":null,"authors":["Fang Wan","Chaoyang Song"],"abstract":"The human reasoning process is seldom a one-way process from an input leading\nto an output. Instead, it often involves a systematic deduction by ruling out\nother possible outcomes as a self-checking mechanism. In this paper, we\ndescribe the design of a hybrid neural network for logical learning that is\nsimilar to the human reasoning through the introduction of an auxiliary input,\nnamely the indicators, that act as the hints to suggest logical outcomes. We\ngenerate these indicators by digging into the hidden information buried\nunderneath the original training data for direct or indirect suggestions. We\nused the MNIST data to demonstrate the design and use of these indicators in a\nconvolutional neural network. We trained a series of such hybrid neural\nnetworks with variations of the indicators. Our results show that these hybrid\nneural networks are very robust in generating logical outcomes with inherently\nhigher prediction accuracy than the direct use of the original input and output\nin apparent models. Such improved predictability with reassured logical\nconfidence is obtained through the exhaustion of all possible indicators to\nrule out all illogical outcomes, which is not available in the apparent models.\nOur logical learning process can effectively cope with the unknown unknowns\nusing a full exploitation of all existing knowledge available for learning. The\ndesign and implementation of the hints, namely the indicators, become an\nessential part of artificial intelligence for logical learning. We also\nintroduce an ongoing application setup for this hybrid neural network in an\nautonomous grasping robot, namely as_DeepClaw, aiming at learning an optimized\ngrasping pose through logical learning.","url_abs":"http://arxiv.org/abs/1705.08200v1","url_pdf":"http://arxiv.org/pdf/1705.08200v1.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":"logical-learning-through-a-hybrid-neural","repo_url":"https://github.com/as-wanfang/as_HybridNN","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","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}