{"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/efficient-sequence-labeling-with-actor-critic","title":"Efficient Sequence Labeling with Actor-Critic Training","arxiv_id":"1810.00428","date":"2018-09-30","proceeding":null,"authors":["Saeed Najafi","Colin Cherry","Grzegorz Kondrak"],"abstract":"Neural approaches to sequence labeling often use a Conditional Random Field\n(CRF) to model their output dependencies, while Recurrent Neural Networks (RNN)\nare used for the same purpose in other tasks. We set out to establish RNNs as\nan attractive alternative to CRFs for sequence labeling. To do so, we address\none of the RNN's most prominent shortcomings, the fact that it is not exposed\nto its own errors with the maximum-likelihood training. We frame the prediction\nof the output sequence as a sequential decision-making process, where we train\nthe network with an adjusted actor-critic algorithm (AC-RNN). We\ncomprehensively compare this strategy with maximum-likelihood training for both\nRNNs and CRFs on three structured-output tasks. The proposed AC-RNN efficiently\nmatches the performance of the CRF on NER and CCG tagging, and outperforms it\non Machine Transliteration. We also show that our training strategy is\nsignificantly better than other techniques for addressing RNN's exposure bias,\nsuch as Scheduled Sampling, and Self-Critical policy training.","url_abs":"http://arxiv.org/abs/1810.00428v1","url_pdf":"http://arxiv.org/pdf/1810.00428v1.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":"efficient-sequence-labeling-with-actor-critic","repo_url":"https://github.com/SaeedNajafi/ac-tagger","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"decision-making","task_name":"Decision Making"},{"task_slug":"cg","task_name":"NER"},{"task_slug":"sequential-decision-making","task_name":"Sequential Decision Making"},{"task_slug":"transliteration","task_name":"Transliteration"}],"methods":[{"method_slug":"crf","method_name":"CRF"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}