As machine learning (ML) systems emerge in end-user applications, learning algorithms and classifiers will need to be robust to an increasingly unpredictable operating environment. In many cases, the parameters governing a learning system cannot be optimized for every user scenario, nor can users typically manipulate parameters defined in the space and terminology of ML. Conventional approaches to user-oriented ML systems have typically hidden this complexity from users by automating parameter adjustment. We propose a new paradigm, in which model and algorithm parameters are exposed directly to end-users with intuitive labels, suitable for applications where parameters cannot be automatically optimized or where there is additional motivation \u2013 such as creative flexibility \u2013 to expose, rather than fix or automatically adapt, learning parameters. In our CHI 2008 paper, we introduced and evaluated MySong, a system that uses a Hidden Markov Model to generate chords to accompany a vocal melody. The present paper formally describes the learning underlying MySong and discusses the mechanisms by which MySong’s learning parameters are exposed to users, as a case study in making ML systems user-configurable. We discuss the generalizability of this approach, and propose that intuitively exposing ML parameters is a key challenge for the ML and human-computer-interaction communities.<\/p>\n","protected":false},"excerpt":{"rendered":"
As machine learning (ML) systems emerge in end-user applications, learning algorithms and classifiers will need to be robust to an increasingly unpredictable operating environment. In many cases, the parameters governing a learning system cannot be optimized for every user scenario, nor can users typically manipulate parameters defined in the space and terminology of ML. Conventional […]<\/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":"ACM","msr_publisher_other":"","msr_booktitle":"","msr_chapter":"","msr_edition":"AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 2","msr_editors":"","msr_how_published":"","msr_isbn":"978-1-57735-368-3","msr_issue":"","msr_journal":"","msr_number":"","msr_organization":"","msr_pages_string":"784-791","msr_page_range_start":"784","msr_page_range_end":"791","msr_series":"","msr_volume":"","msr_copyright":"","msr_conference_name":"AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 2","msr_doi":"","msr_arxiv_id":"","msr_s2_paper_id":"","msr_mag_id":"","msr_pubmed_id":"","msr_other_authors":"Ian Simon","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":"2008-07-13","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":2008,"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":0,"footnotes":""},"msr-research-highlight":[],"research-area":[13554],"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-145399","msr-research-item","type-msr-research-item","status-publish","hentry","msr-research-area-human-computer-interaction","msr-locale-en_us"],"msr_publishername":"ACM","msr_edition":"AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 2","msr_affiliation":"","msr_published_date":"2008-07-13","msr_host":"","msr_duration":"","msr_version":"","msr_speaker":"","msr_other_contributors":"","msr_booktitle":"","msr_pages_string":"784-791","msr_chapter":"","msr_isbn":"978-1-57735-368-3","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":"208410","msr_publicationurl":"","msr_doi":"","msr_publication_uploader":[{"type":"file","title":"mysongaaai2008.pdf","viewUrl":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2016\/02\/mysongaaai2008.pdf","id":208410,"label_id":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":[{"id":208410,"url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-content\/uploads\/2016\/02\/mysongaaai2008.pdf"}],"msr-author-ordering":[{"type":"user_nicename","value":"dan","user_id":31522,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=dan"},{"type":"text","value":"Ian Simon","user_id":0,"rest_url":false},{"type":"user_nicename","value":"sumitb","user_id":33754,"rest_url":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/microsoft-research\/v1\/researchers?person=sumitb"}],"msr_impact_theme":[],"msr_research_lab":[199565],"msr_event":[],"msr_group":[],"msr_project":[327332,326717],"publication":[],"video":[],"msr-tool":[],"msr_publication_type":"inproceedings","related_content":{"projects":[{"ID":327332,"post_title":"Computational Tools for Music","post_name":"computational-tools-for-music","post_type":"msr-project","post_date":"2016-11-27 16:38:01","post_modified":"2021-05-09 12:03:16","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/computational-tools-for-music\/","post_excerpt":"Work in this area seeks to use computational tools to enable musical creativity, in particular to give novices a variety of new approaches to experience musical creativity.","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/327332"}]}},{"ID":326717,"post_title":"MySong: Automatic Accompaniment for Vocal Melodies","post_name":"mysong-automatic-accompaniment-vocal-melodies","post_type":"msr-project","post_date":"2016-11-23 16:06:54","post_modified":"2021-05-09 12:03:26","post_status":"publish","permalink":"https:\/\/www.microsoft.com\/en-us\/research\/project\/mysong-automatic-accompaniment-vocal-melodies\/","post_excerpt":"Want to Give MySong a Try? MySong is now Songsmith; you can download a free trial (or the full version, if you\u2019re a teacher who wants to use it in your classroom) at: http:\/\/research.microsoft.com\/songsmith Happy Songsmith\u2019ing! Like to Write Music? Most folks never get a chance to answer this question, since writing music takes years of experience... if you don\u2019t play an instrument or spend lots of time around music, you\u2019ll probably never get to…","_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/326717"}]}}]},"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/145399","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-research-item"}],"version-history":[{"count":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/145399\/revisions"}],"predecessor-version":[{"id":524746,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-item\/145399\/revisions\/524746"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=145399"}],"wp:term":[{"taxonomy":"msr-research-highlight","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-research-highlight?post=145399"},{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=145399"},{"taxonomy":"msr-publication-type","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publication-type?post=145399"},{"taxonomy":"msr-publisher","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-publisher?post=145399"},{"taxonomy":"msr-focus-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-focus-area?post=145399"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=145399"},{"taxonomy":"msr-post-option","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-post-option?post=145399"},{"taxonomy":"msr-field-of-study","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-field-of-study?post=145399"},{"taxonomy":"msr-conference","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-conference?post=145399"},{"taxonomy":"msr-journal","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-journal?post=145399"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=145399"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=145399"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}