{"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/rethinking-action-spaces-for-reinforcement","title":"Rethinking Action Spaces for Reinforcement Learning in End-to-end Dialog Agents with Latent Variable Models","arxiv_id":"1902.08858","date":"2019-02-23","proceeding":"NAACL 2019 6","authors":["Tiancheng Zhao","Kaige Xie","Maxine Eskenazi"],"abstract":"Defining action spaces for conversational agents and optimizing their\ndecision-making process with reinforcement learning is an enduring challenge.\nCommon practice has been to use handcrafted dialog acts, or the output\nvocabulary, e.g. in neural encoder decoders, as the action spaces. Both have\ntheir own limitations. This paper proposes a novel latent action framework that\ntreats the action spaces of an end-to-end dialog agent as latent variables and\ndevelops unsupervised methods in order to induce its own action space from the\ndata. Comprehensive experiments are conducted examining both continuous and\ndiscrete action types and two different optimization methods based on\nstochastic variational inference. Results show that the proposed latent actions\nachieve superior empirical performance improvement over previous word-level\npolicy gradient methods on both DealOrNoDeal and MultiWoz dialogs. Our detailed\nanalysis also provides insights about various latent variable approaches for\npolicy learning and can serve as a foundation for developing better latent\nactions in future research.","url_abs":"http://arxiv.org/abs/1902.08858v2","url_pdf":"http://arxiv.org/pdf/1902.08858v2.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":"rethinking-action-spaces-for-reinforcement","repo_url":"https://github.com/snakeztc/NeuralDialog-LaRL","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"rethinking-action-spaces-for-reinforcement","repo_url":"https://github.com/Jupaoqq/Jupaoqq_LaRL","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}},{"paper_slug":"rethinking-action-spaces-for-reinforcement","repo_url":"https://github.com/justinchiu/NeuralDialog","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"decision-making","task_name":"Decision Making"},{"task_slug":"dialogue-generation","task_name":"Dialogue Generation"},{"task_slug":"dialogue-management","task_name":"Dialogue Management"},{"task_slug":"goal-oriented-dialogue-systems","task_name":"Goal-Oriented Dialogue Systems"},{"task_slug":"policy-gradient-methods","task_name":"Policy Gradient Methods"},{"task_slug":"reinforcement-learning","task_name":"Reinforcement Learning"},{"task_slug":"variational-inference","task_name":"Variational Inference"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1902.08858","atlas_url":"https://app.syntology.ai/?focus=1902.08858","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"1902.08858"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-25T09:33:49+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/snakeztc/NeuralDialog-LaRL","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/justinchiu/NeuralDialog","reach":{"status":"ok","spdx":"Apache-2.0"}},{"provenance":"external:paperswithcode_snapshot_2025-07-28","url":"https://github.com/Jupaoqq/Jupaoqq_LaRL","reach":{"status":"ok","spdx":"Apache-2.0"}}],"summary":{"unverified":6},"by_repo_kind":{"official":{"samples":6,"ran":0,"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":"dc0b614a9138b9f0","entry":"backward_hook","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/utils.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"dc0b614a9138b9f0"}},{"code_sha256_prefix":"0b8bbebd022d579a","entry":"get_domain","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/domain.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/domain.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"0b8bbebd022d579a"}},{"code_sha256_prefix":"ecb6dfa50e44912b","entry":"get_tag","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/data.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/data.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"ecb6dfa50e44912b"}},{"code_sha256_prefix":"7f7e4b0b6f9f8f19","entry":"load_model","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/utils.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"7f7e4b0b6f9f8f19"}},{"code_sha256_prefix":"4f6d65cb0a38293a","entry":"read_lines","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/data.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/data.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"4f6d65cb0a38293a"}},{"code_sha256_prefix":"6990ccd38a23be9f","entry":"use_cuda","repo":"snakeztc/NeuralDialog-LaRL","repo_kind":"official","path":"FB/utils.py","file_url":"https://github.com/snakeztc/NeuralDialog-LaRL/blob/HEAD/FB/utils.py","link_basis":"first_harvest_node","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"Apache-2.0","inline_ok":true,"mcp_get_code":{"code_sha256":"6990ccd38a23be9f"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}