{"id":545241,"date":"2018-11-20T10:10:06","date_gmt":"2018-11-20T18:10:06","guid":{"rendered":"https:\/\/www.microsoft.com\/en-us\/research\/?post_type=msr-project&p=545241"},"modified":"2022-10-27T11:29:20","modified_gmt":"2022-10-27T18:29:20","slug":"automl","status":"publish","type":"msr-project","link":"https:\/\/www.microsoft.com\/en-us\/research\/project\/automl\/","title":{"rendered":"AutoML"},"content":{"rendered":"

State-of-the-art machine learning\/AI systems consist of complex pipelines with choices of hyperparameters, models and configuration details that need to be tuned for optimal performance. The resulting optimization space can be too complex and high-dimensional for researchers and engineers to explore manually. When automated systems are used, the high costs of running a single experiment (e.g. training a deep neural network) and the high sample complexity (i.e. large number of experiments required) together make na\u00efve approaches impractical.<\/p>\n

Many of the problems we are interested in can be cast as high-dimensional combinatorial optimization tasks.\u00a0 Broadly speaking, we tackle these problems by designing probabilistic machine learning models to guide (automated) experimental decisions and meta-learning to reduce the sample complexity and transfer knowledge across related datasets or problems.<\/p>\n

Specific problems that the Microsoft Research AutoML team focuses on include:<\/p>\n\n\n\n\n\n\n\n
\"\"<\/td>\nNeural architecture search<\/td>\n<\/tr>\n
\"\"<\/td>\nModel selection<\/td>\n<\/tr>\n
\"\"<\/td>\nFeature engineering<\/td>\n<\/tr>\n
\"\"<\/td>\nHyperparameter tuning<\/td>\n<\/tr>\n
\"\"<\/td>\nModel compression<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

 <\/p>\n

Core AutoML technology is already live in Azure Machine Learning (opens in new tab)<\/span><\/a>, Power BI (opens in new tab)<\/span><\/a> and other Microsoft products.<\/p>\n

 <\/p>\nOpens in a new tab<\/span>","protected":false},"excerpt":{"rendered":"

State-of-the-art machine learning\/AI systems consist of complex pipelines with choices of hyperparameters, models and configuration details that need to be tuned for optimal performance. The AutoML project develops automated methods for optimizing AI pipelines, making AI development more broadly accessible. Core AutoML technology is already live in Azure Machine Learning, Power BI and other Microsoft products. Our AutoML research is advancing the state of the art in neural architecture search, model compression and more.<\/p>\n","protected":false},"featured_media":599691,"template":"","meta":{"msr-url-field":"","msr-podcast-episode":"","msrModifiedDate":"","msrModifiedDateEnabled":false,"ep_exclude_from_search":false,"footnotes":""},"research-area":[13556],"msr-impact-theme":[],"msr-pillar":[],"msr_project_start":"","related-publications":[652095,692796,715519,732229,764053,507689,647112],"related-downloads":[713053],"related-videos":[],"related-groups":[],"related-events":[],"related-opportunities":[],"related-posts":[],"related-articles":[],"tab-content":[],"slides":[{"attachment_id":674697,"headline":"Directions in ML: AutoML virtual speaker series","cta":"Save your virtual seat and register for the upcoming talk","url":"https:\/\/www.microsoft.com\/en-us\/research\/event\/directions-in-ml\/","cta_style":"","slideshow_type":"feature"}],"related-researchers":[{"type":"user_nicename","display_name":"Nicolo Fusi","user_id":31829,"people_section":"Team","alias":"fusi"},{"type":"user_nicename","display_name":"Sharon Gillett","user_id":33599,"people_section":"Team","alias":"sharong"},{"type":"user_nicename","display_name":"Jimmy Hall","user_id":38148,"people_section":"Team","alias":"jamhall"},{"type":"user_nicename","display_name":"Neil Tenenholtz","user_id":38464,"people_section":"Team","alias":"netenenh"},{"type":"user_nicename","display_name":"Lester Mackey","user_id":36161,"people_section":"Team","alias":"lmackey"},{"type":"user_nicename","display_name":"Philip Rosenfield","user_id":37562,"people_section":"Team","alias":"phrosenf"},{"type":"user_nicename","display_name":"David Alvarez-Melis","user_id":38814,"people_section":"Team","alias":"daalvare"}],"msr_research_lab":[199563],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/545241"}],"collection":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project"}],"about":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/types\/msr-project"}],"version-history":[{"count":26,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/545241\/revisions"}],"predecessor-version":[{"id":786922,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/545241\/revisions\/786922"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media\/599691"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=545241"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=545241"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=545241"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=545241"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}