{"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/latent-alignment-and-variational-attention","title":"Latent Alignment and Variational Attention","arxiv_id":"1807.03756","date":"2018-07-10","proceeding":"NeurIPS 2018 12","authors":["Yuntian Deng","Yoon Kim","Justin Chiu","Demi Guo","Alexander M. Rush"],"abstract":"Neural attention has become central to many state-of-the-art models in\nnatural language processing and related domains. Attention networks are an\neasy-to-train and effective method for softly simulating alignment; however,\nthe approach does not marginalize over latent alignments in a probabilistic\nsense. This property makes it difficult to compare attention to other alignment\napproaches, to compose it with probabilistic models, and to perform posterior\ninference conditioned on observed data. A related latent approach, hard\nattention, fixes these issues, but is generally harder to train and less\naccurate. This work considers variational attention networks, alternatives to\nsoft and hard attention for learning latent variable alignment models, with\ntighter approximation bounds based on amortized variational inference. We\nfurther propose methods for reducing the variance of gradients to make these\napproaches computationally feasible. Experiments show that for machine\ntranslation and visual question answering, inefficient exact latent variable\nmodels outperform standard neural attention, but these gains go away when using\nhard attention based training. On the other hand, variational attention retains\nmost of the performance gain but with training speed comparable to neural\nattention.","url_abs":"http://arxiv.org/abs/1807.03756v2","url_pdf":"http://arxiv.org/pdf/1807.03756v2.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":"latent-alignment-and-variational-attention","repo_url":"https://github.com/harvardnlp/var-attn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"hard-attention","task_name":"Hard Attention"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"translation","task_name":"Translation"},{"task_slug":"variational-inference","task_name":"Variational Inference"},{"task_slug":"visual-question-answering-1","task_name":"Visual Question Answering"},{"task_slug":"visual-question-answering","task_name":"Visual Question Answering (VQA)"}],"methods":[{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/machine-translation-on-iwslt2014-german","task":"Machine Translation","dataset":"IWSLT2014 German-English","model":"Variational Attention","rank_in_archive_order":29,"of":34,"metrics":{"BLEU score":"33.1"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.03756","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1807.03756"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/harvardnlp/var-attn","reach":null}],"summary":{"ran_fixture":1,"unverified":1},"by_repo_kind":{"official":{"samples":2,"ran":1,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"6aefb46aa147be92","entry":"sample_gumbel","repo":"harvardnlp/var-attn","repo_kind":"official","path":"onmt/modules/VariationalAttention.py","file_url":"https://github.com/harvardnlp/var-attn/blob/HEAD/onmt/modules/VariationalAttention.py","link_basis":"first_harvest_node","language":"python","status":"ran_fixture","verification_level":1,"contract_check":"RAISES","metamorphic_tier":"invariant","behaviour_fingerprint":true,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"6aefb46aa147be92"}},{"code_sha256_prefix":"1fca2249a86bd6ac","entry":"gumbel_softmax_sample","repo":"harvardnlp/var-attn","repo_kind":"official","path":"onmt/modules/VariationalAttention.py","file_url":"https://github.com/harvardnlp/var-attn/blob/HEAD/onmt/modules/VariationalAttention.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1fca2249a86bd6ac"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}