{"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/probabilistic-reasoning-via-deep-learning","title":"Probabilistic Reasoning via Deep Learning: Neural Association Models","arxiv_id":"1603.07704","date":"2016-03-24","proceeding":null,"authors":["Quan Liu","Hui Jiang","Andrew Evdokimov","Zhen-Hua Ling","Xiaodan Zhu","Si Wei","Yu Hu"],"abstract":"In this paper, we propose a new deep learning approach, called neural\nassociation model (NAM), for probabilistic reasoning in artificial\nintelligence. We propose to use neural networks to model association between\nany two events in a domain. Neural networks take one event as input and compute\na conditional probability of the other event to model how likely these two\nevents are to be associated. The actual meaning of the conditional\nprobabilities varies between applications and depends on how the models are\ntrained. In this work, as two case studies, we have investigated two NAM\nstructures, namely deep neural networks (DNN) and relation-modulated neural\nnets (RMNN), on several probabilistic reasoning tasks in AI, including\nrecognizing textual entailment, triple classification in multi-relational\nknowledge bases and commonsense reasoning. Experimental results on several\npopular datasets derived from WordNet, FreeBase and ConceptNet have all\ndemonstrated that both DNNs and RMNNs perform equally well and they can\nsignificantly outperform the conventional methods available for these reasoning\ntasks. Moreover, compared with DNNs, RMNNs are superior in knowledge transfer,\nwhere a pre-trained model can be quickly extended to an unseen relation after\nobserving only a few training samples. To further prove the effectiveness of\nthe proposed models, in this work, we have applied NAMs to solving challenging\nWinograd Schema (WS) problems. Experiments conducted on a set of WS problems\nprove that the proposed models have the potential for commonsense reasoning.","url_abs":"http://arxiv.org/abs/1603.07704v2","url_pdf":"http://arxiv.org/pdf/1603.07704v2.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":[],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":"natural-language-inference","task_name":"Natural Language Inference"},{"task_slug":"natural-language-understanding","task_name":"Natural Language Understanding"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"triple-classification","task_name":"Triple Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/natural-language-understanding-on-pdp60","task":"Natural Language Understanding","dataset":"PDP60","model":"USSM + Cause-Effect Knowledge Base","rank_in_archive_order":11,"of":13,"metrics":{"Accuracy":"55.0"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1603.07704","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}