{"id":656004,"date":"2020-05-04T13:20:17","date_gmt":"2020-05-04T20:20:17","guid":{"rendered":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/?post_type=msr-research-item&#038;p=656004"},"modified":"2025-09-02T07:32:12","modified_gmt":"2025-09-02T14:32:12","slug":"fast-domain-adaptation-for-goal-oriented-dialogue-using-a-hybrid-generative-retrieval-transformer","status":"publish","type":"msr-research-item","link":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/publication\/fast-domain-adaptation-for-goal-oriented-dialogue-using-a-hybrid-generative-retrieval-transformer\/","title":{"rendered":"Fast Domain Adaptation For Goal-Oriented Dialogue Using A Hybrid Generative-Retrieval Transformer"},"content":{"rendered":"<p>Goal-oriented dialogue systems are now widely adopted in industry, where practical aspects of using them becomes of key importance. As such, it is expected from such systems to fit into a rapid prototyping cycle for new products and domains. For data-driven dialogue systems (especially those based on deep learning) that amounts to maintaining production-level performance having been provided with a few \u2018seed\u2019 dialogue examples, normally referred to as data efficiency. With extremely data-dependent deep learning methods, the most promising way to achieve practical data efficiency is transfer learning\u2014i.e., leveraging a greater, highly represented data source for training a base model, then fine-tuning it to available in-domain data. In this paper, we present a hybrid generative-retrieval model that can be trained using transfer learning. By using GPT-2 as the base model and fine-tuning it to the multidomain MetaLWOz dataset, we obtain a robust dialogue model able to perform both response generation and ranking <sup>1<\/sup>\u00a0. Combining both, it outperforms several competitive generative-only and retrieval-only baselines, measured by language modeling quality on MetaLWOz as well as in goal- oriented metrics (Intent\/Slot Fl-scores) on the MultiWoz corpus.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Goal-oriented dialogue systems are now widely adopted in industry, where practical aspects of using them becomes of key importance. As such, it is expected from such systems to fit into a rapid prototyping cycle for new products and domains. For data-driven dialogue systems (especially those based on deep learning) that amounts to maintaining production-level performance [&hellip;]<\/p>\n","protected":false},"featured_media":0,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"_classifai_error":"","msr-author-ordering":null,"msr_publishername":"","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"","msr_editors":"","msr_how_published":"","msr_isbn":"","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"","msr_page_range_start":"","msr_page_range_end":"","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","msr_doi":"","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"","msr_pubmed_id":"","msr_other_authors":"","msr_other_contributors":"","msr_speaker":"","msr_award":"","msr_affiliation":"","msr_institution":"","msr_host":"","msr_version":"","msr_duration":"","msr_original_fields_of_study":"","msr_release_tracker_id":"","msr_s2_match_type":"","msr_citation_count_updated":"","msr_published_date":"2020-4-9","msr_highlight_text":"","msr_notes":"","msr_longbiography":"","msr_publicationurl":"","msr_external_url":"","msr_secondary_video_url":"","msr_conference_url":"","msr_journal_url":"","msr_s2_pdf_url":"","msr_year":0,"msr_citation_count":0,"msr_influential_citations":0,"msr_reference_count":0,"msr_s2_match_confidence":0,"msr_microsoftintellectualproperty":true,"msr_s2_open_access":false,"msr_s2_author_ids":[],"msr_pub_ids":[],"msr_hide_image_in_river":null,"footnotes":""},"msr-research-highlight":[],"research-area":[243062],"msr-publication-type":[193716],"msr-publisher":[],"msr-focus-area":[],"msr-locale":[268875],"msr-post-option":[],"msr-field-of-study":[],"msr-conference":[],"msr-journal":[],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-656004","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-audio-acoustics","msr-locale-en_us"],"msr_publishername":"","msr_edition":"","msr_affiliation":"","msr_published_date":"2020-4-9","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"","msr_chapter":"","msr_isbn":"","msr_journal":"","msr_volume":"","msr_number":"","msr_editors":"","msr_series":"","msr_issue":"","msr_organization":"","msr_how_published":"","msr_notes":"","msr_highlight_text":"","msr_release_tracker_id":"","msr_original_fields_of_study":"","msr_download_urls":"","msr_external_url":"","msr_secondary_video_url":"","msr_longbiography":"","msr_microsoftintellectualproperty":1,"msr_main_download":"","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"url","viewUrl":"false","id":"false","title":"https:\/\/ieeexplore.ieee.org\/abstract\/document\/9053599","label_id":"243109","label":0}],"msr_related_uploader":"","msr_citation_count":0,"msr_citation_count_updated":"","msr_s2_paper_id":"","msr_influential_citations":0,"msr_reference_count":0,"msr_arxiv_id":"","msr_s2_author_ids":[],"msr_s2_open_access":false,"msr_s2_pdf_url":null,"msr_attachments":[],"msr-author-ordering":[{"type":"text","value":"Igor Shalyminov","user_id":0,"rest_url":false},{"type":"user_nicename","value":"Alessandro Sordoni","user_id":37230,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Alessandro Sordoni"},{"type":"user_nicename","value":"Adam Atkinson","user_id":37095,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Adam Atkinson"},{"type":"user_nicename","value":"Hannes Schulz","user_id":37188,"rest_url":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=Hannes Schulz"}],"msr_impact_theme":[],"msr_research_lab":[437514],"msr_event":[],"msr_group":[896463,1148823],"msr_project":[583807],"publication":[],"video":[],"msr-tool":[600426],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":583807,"post_title":"MetaLWOz","post_name":"metalwoz","post_type":"msr-project","post_date":"2019-06-14 08:40:23","post_modified":"2019-12-12 06:47:20","post_status":"publish","permalink":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/project\/metalwoz\/","post_excerpt":"Meta-Learning Wizard-of-Oz (MetaLWOz) is a dataset designed to help develop models capable of predicting user responses in unseen domains.","_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/583807"}]}}]},"_links":{"self":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/656004","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":2,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/656004\/revisions"}],"predecessor-version":[{"id":1148992,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/656004\/revisions\/1148992"}],"wp:attachment":[{"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/media?parent=656004"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=656004"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=656004"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=656004"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=656004"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=656004"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=656004"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=656004"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=656004"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=656004"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=656004"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=656004"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/cm-edgetun.pages.dev\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=656004"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}