{"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/attentive-convolution-equipping-cnns-with-rnn","title":"Attentive Convolution: Equipping CNNs with RNN-style Attention Mechanisms","arxiv_id":"1710.00519","date":"2017-10-02","proceeding":"TACL 2018 1","authors":["Wenpeng Yin","Hinrich Schütze"],"abstract":"In NLP, convolutional neural networks (CNNs) have benefited less than\nrecurrent neural networks (RNNs) from attention mechanisms. We hypothesize that\nthis is because the attention in CNNs has been mainly implemented as attentive\npooling (i.e., it is applied to pooling) rather than as attentive convolution\n(i.e., it is integrated into convolution). Convolution is the differentiator of\nCNNs in that it can powerfully model the higher-level representation of a word\nby taking into account its local fixed-size context in the input text t^x. In\nthis work, we propose an attentive convolution network, ATTCONV. It extends the\ncontext scope of the convolution operation, deriving higher-level features for\na word not only from local context, but also information extracted from\nnonlocal context by the attention mechanism commonly used in RNNs. This\nnonlocal context can come (i) from parts of the input text t^x that are distant\nor (ii) from extra (i.e., external) contexts t^y. Experiments on sentence\nmodeling with zero-context (sentiment analysis), single-context (textual\nentailment) and multiple-context (claim verification) demonstrate the\neffectiveness of ATTCONV in sentence representation learning with the\nincorporation of context. In particular, attentive convolution outperforms\nattentive pooling and is a strong competitor to popular attentive RNNs.","url_abs":"http://arxiv.org/abs/1710.00519v2","url_pdf":"http://arxiv.org/pdf/1710.00519v2.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":"attentive-convolution-equipping-cnns-with-rnn","repo_url":"https://github.com/kenkenling/NLP","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"claim-verification","task_name":"Claim Verification"},{"task_slug":"natural-language-inference","task_name":"Natural Language Inference"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"sentiment-analysis","task_name":"Sentiment Analysis"}],"methods":[{"method_slug":"convolution","method_name":"Convolution"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1710.00519","atlas_url":"https://app.syntology.ai/?focus=1710.00519","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}