Reinforcement learning (RL) has achieved great success in video and board games. In this project, we aim at boosting machine learning algorithms and systems by leveraging reinforcement learning techniques. We focus the following aspects. First, RL for data selection and pre-processing, in which we use RL techniques to select right data at right time and process the data in a right way for model training. Second, RL for hyper parameter optimization. Setting appropriate hyper parameters is important for learning algorithms. We use RL techniques to optimize hyper parameters for deep algorithms, including learning rate, gradients, momentum, \u2026 Third, RL for deep structure optimization. Designing a good structure is critical for applications, such as CNN for image related tasks and RNN for sequence related tasks. We leverage RL techniques to find and design better deep structures for practical applications.<\/p>\n","protected":false},"excerpt":{"rendered":"
Reinforcement learning (RL) has achieved great success in video and board games. In this project, we aim at boosting machine learning algorithms and systems by leveraging reinforcement learning techniques. We focus the following aspects. First, RL for data selection and pre-processing, in which we use RL techniques to select right data at right time and […]<\/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":"","footnotes":""},"research-area":[13556],"msr-locale":[268875],"msr-impact-theme":[],"msr-pillar":[],"class_list":["post-272661","msr-project","type-msr-project","status-publish","hentry","msr-research-area-artificial-intelligence","msr-locale-en_us","msr-archive-status-active"],"msr_project_start":"2016-08-01","related-publications":[],"related-downloads":[],"related-videos":[],"related-groups":[],"related-events":[],"related-opportunities":[],"related-posts":[],"related-articles":[],"tab-content":[],"slides":[],"related-researchers":[{"type":"user_nicename","value":"taoqin","display_name":"Tao Qin","author_link":"Tao Qin<\/a>","is_active":false,"user_id":33871,"last_first":"Qin, Tao","people_section":0,"alias":"taoqin"},{"type":"user_nicename","value":"tyliu","display_name":"Tie-Yan Liu","author_link":"Tie-Yan Liu<\/a>","is_active":false,"user_id":34431,"last_first":"Liu, Tie-Yan","people_section":0,"alias":"tyliu"}],"msr_research_lab":[199560],"msr_impact_theme":[],"_links":{"self":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/272661"}],"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":1,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/272661\/revisions"}],"predecessor-version":[{"id":390527,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-project\/272661\/revisions\/390527"}],"wp:attachment":[{"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/media?parent=272661"}],"wp:term":[{"taxonomy":"msr-research-area","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/research-area?post=272661"},{"taxonomy":"msr-locale","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-locale?post=272661"},{"taxonomy":"msr-impact-theme","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-impact-theme?post=272661"},{"taxonomy":"msr-pillar","embeddable":true,"href":"https:\/\/www.microsoft.com\/en-us\/research\/wp-json\/wp\/v2\/msr-pillar?post=272661"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}