{"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/towards-neural-network-based-reasoning","title":"Towards Neural Network-based Reasoning","arxiv_id":"1508.05508","date":"2015-08-22","proceeding":null,"authors":["Baolin Peng","Zhengdong Lu","Hang Li","Kam-Fai Wong"],"abstract":"We propose Neural Reasoner, a framework for neural network-based reasoning\nover natural language sentences. Given a question, Neural Reasoner can infer\nover multiple supporting facts and find an answer to the question in specific\nforms. Neural Reasoner has 1) a specific interaction-pooling mechanism,\nallowing it to examine multiple facts, and 2) a deep architecture, allowing it\nto model the complicated logical relations in reasoning tasks. Assuming no\nparticular structure exists in the question and facts, Neural Reasoner is able\nto accommodate different types of reasoning and different forms of language\nexpressions. Despite the model complexity, Neural Reasoner can still be trained\neffectively in an end-to-end manner. Our empirical studies show that Neural\nReasoner can outperform existing neural reasoning systems with remarkable\nmargins on two difficult artificial tasks (Positional Reasoning and Path\nFinding) proposed in [8]. For example, it improves the accuracy on Path\nFinding(10K) from 33.4% [6] to over 98%.","url_abs":"http://arxiv.org/abs/1508.05508v1","url_pdf":"http://arxiv.org/pdf/1508.05508v1.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":"towards-neural-network-based-reasoning","repo_url":"https://github.com/rgsachin/DMTN","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"MIT"}}],"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}