{"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/neural-tree-indexers-for-text-understanding","title":"Neural Tree Indexers for Text Understanding","arxiv_id":"1607.04492","date":"2016-07-15","proceeding":"EACL 2017 4","authors":["Tsendsuren Munkhdalai","Hong Yu"],"abstract":"Recurrent neural networks (RNNs) process input text sequentially and model\nthe conditional transition between word tokens. In contrast, the advantages of\nrecursive networks include that they explicitly model the compositionality and\nthe recursive structure of natural language. However, the current recursive\narchitecture is limited by its dependence on syntactic tree. In this paper, we\nintroduce a robust syntactic parsing-independent tree structured model, Neural\nTree Indexers (NTI) that provides a middle ground between the sequential RNNs\nand the syntactic treebased recursive models. NTI constructs a full n-ary tree\nby processing the input text with its node function in a bottom-up fashion.\nAttention mechanism can then be applied to both structure and node function. We\nimplemented and evaluated a binarytree model of NTI, showing the model achieved\nthe state-of-the-art performance on three different NLP tasks: natural language\ninference, answer sentence selection, and sentence classification,\noutperforming state-of-the-art recurrent and recursive neural networks.","url_abs":"http://arxiv.org/abs/1607.04492v2","url_pdf":"http://arxiv.org/pdf/1607.04492v2.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":"neural-tree-indexers-for-text-understanding","repo_url":"https://bitbucket.org/tsendeemts/nti","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"natural-language-inference","task_name":"Natural Language Inference"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"sentence-classification","task_name":"Sentence Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/natural-language-inference-on-snli","task":"Natural Language Inference","dataset":"SNLI","model":"300D Full tree matching NTI-SLSTM-LSTM w/ global attention","rank_in_archive_order":46,"of":98,"metrics":{"% Test Accuracy":"87.3","% Train Accuracy":"88.5","Parameters":"3.2m"},"uses_additional_data":false},{"leaderboard":"/sota/natural-language-inference-on-snli","task":"Natural Language Inference","dataset":"SNLI","model":"300D NTI-SLSTM-LSTM encoders","rank_in_archive_order":84,"of":98,"metrics":{"% Test Accuracy":"83.4","% Train Accuracy":"82.5","Parameters":"4.0m"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1607.04492","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}