{"639515":{"#nid":"639515","#data":{"type":"event","title":"PhD Defense by Fred Hohman","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ETitle:\u003C\/strong\u003E\u0026nbsp;Interactive Scalable Interfaces for Machine Learning Interpretability\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003EFred Hohman\u003C\/strong\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Ca href=\u0022http:\/\/fredhohman.com\u0022\u003Ehttps:\/\/fredhohman.com\/\u003C\/a\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003ESchool of Computational Science \u0026amp; Engineering\u003Cbr \/\u003E\r\nCollege of Computing\u003Cbr \/\u003E\r\nGeorgia Institute of Technology\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003EDate:\u003C\/strong\u003E\u0026nbsp;Monday, October 5, 2020\u003Cbr \/\u003E\r\n\u003Cstrong\u003ETime:\u003C\/strong\u003E\u0026nbsp;1:30pm - 2:30pm EST\u003Cbr \/\u003E\r\n\u003Cstrong\u003ELocation\u003C\/strong\u003E\u0026nbsp;(remote via Zoom, password: fred):\u0026nbsp;\u003Ca href=\u0022https:\/\/us02web.zoom.us\/j\/89967814013?pwd=T0RaeEJMSEt4YndsR1VkbGFVcFo1UT09\u0022\u003Ehttps:\/\/us02web.zoom.us\/j\/89967814013?pwd=T0RaeEJMSEt4YndsR1VkbGFVcFo1UT09\u003C\/a\u003E\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003ECommittee\u003C\/strong\u003E\u003Cbr \/\u003E\r\nDuen Horng (Polo) Chau - Advisor, Georgia Tech, Computational Science \u0026amp; Engineering\u003Cbr \/\u003E\r\nAlex Endert - Co-advisor, Georgia Tech, Interactive Computing\u003C\/p\u003E\r\n\r\n\u003Cp\u003EChao Zhang - Georgia Tech, Computational Science \u0026amp; Engineering\u003Cbr \/\u003E\r\nNathan Hodas - Pacific Northwest National Lab\u003Cbr \/\u003E\r\nScott Davidoff - NASA Jet Propulsion Lab\u003Cbr \/\u003E\r\nSteven Drucker - Microsoft Research\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cstrong\u003EAbstract\u003C\/strong\u003E\u003Cbr \/\u003E\r\nData-driven paradigms now solve the world\u0026#39;s hardest problems by automatically learning from data.\u0026nbsp;Unfortunately, what is learned is often unknown to both the people who train the models and the people they impact.\u0026nbsp;This has led to a rallying cry for\u0026nbsp;machine learning interpretability.\u0026nbsp;But how do we enable interpretability?\u0026nbsp;How do we scale up explanations for modern, complex models?\u0026nbsp;And how can we best communicate them to people?\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\nSince machine learning now impacts people\u0026#39;s daily lives, we answer these questions taking a\u0026nbsp;\u003Cem\u003Ehuman-centered perspective\u003C\/em\u003E\u0026nbsp;by designing and developing interactive interfaces can enable interpretability at scale and for everyone.\u0026nbsp;This thesis focuses on:\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cem\u003E(1)\u0026nbsp;Enabling machine learning interpretability:\u003C\/em\u003E\u0026nbsp;User research with practitioners guides the creation of our novel operationalization for interpretability, which helps tool builders design interactive systems for model and prediction explanations.\u0026nbsp;We develop two such visualization systems,\u0026nbsp;Gamut\u0026nbsp;and\u0026nbsp;TeleGam,\u0026nbsp;which we deploy at Microsoft Research as a design probe to investigate the emerging practice of interpreting models.\u0026nbsp;\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cem\u003E(2)\u0026nbsp;Scaling deep learning interpretability:\u003C\/em\u003E\u0026nbsp;Our first-of-its-kind\u0026nbsp;Interrogative Survey\u0026nbsp;reveals critical yet understudied areas of deep learning interpretability research, such as the lack of higher-level explanations for neural networks.\u0026nbsp;Through Summit,\u0026nbsp;an interactive visualization system, we present the first scalable graph representation that summarizes and visualizes what features deep learning models learn and how those features interact to make predictions (e.g., InceptionNet trained on ImageNet with 1.2M+ images).\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\n\u003Cem\u003E(3)\u0026nbsp;Communicating interpretability with interactive articles:\u003C\/em\u003E\u0026nbsp;We use interactive articles, a new medium on the web, to teach people about machine learning\u0026#39;s capabilities and limitations, while developing a new interactive publishing initiative called the\u0026nbsp;Parametric Press.\u0026nbsp;From our success publishing interactive content at scale, we generalize and detail the affordances of\u0026nbsp;interactive articles\u0026nbsp;by connecting techniques used in practice and the theories and empirical evaluations put forth by diverse disciplines of research.\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cbr \/\u003E\r\nThis thesis contributes to\u0026nbsp;information visualization,\u0026nbsp;machine learning, and more importantly\u0026nbsp;\u003Cem\u003Etheir intersection\u003C\/em\u003E, including open-source interactive interfaces, scalable algorithms, and new, accessible communication paradigms.\u0026nbsp;Our work is making significant impact in industry and society: our visualizations have been deployed and demoed at Microsoft and built into widely-used interpretability toolkits, our interactive articles have been read by 250,000+ people, and our interpretability research is supported by NASA.\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u003Cstrong\u003EZoom Meeting Invitation\u003C\/strong\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003ETopic: Fred Hohman\u0026#39;s Dissertation Defense\u003Cbr \/\u003E\r\nTime: Oct 5, 2020 01:30 PM Eastern Time (US and Canada)\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\nJoin Zoom Meeting\u003Cbr \/\u003E\r\n\u003Ca href=\u0022https:\/\/us02web.zoom.us\/j\/89967814013?pwd=T0RaeEJMSEt4YndsR1VkbGFVcFo1UT09\u0022\u003Ehttps:\/\/us02web.zoom.us\/j\/89967814013?pwd=T0RaeEJMSEt4YndsR1VkbGFVcFo1UT09\u003C\/a\u003E\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\nMeeting ID: 899 6781 4013\u003Cbr \/\u003E\r\nPasscode: fred\u003Cbr \/\u003E\r\nOne tap mobile\u003Cbr \/\u003E\r\n+13126266799,,89967814013#,,,,,,0#,,374057# US (Chicago)\u003Cbr \/\u003E\r\n+16465588656,,89967814013#,,,,,,0#,,374057# US (New York)\u003Cbr \/\u003E\r\n\u003Cbr \/\u003E\r\nDial by your location\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 312 626 6799 US (Chicago)\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 646 558 8656 US (New York)\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 301 715 8592 US (Germantown)\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 346 248 7799 US (Houston)\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 669 900 9128 US (San Jose)\u003Cbr \/\u003E\r\n\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; +1 253 215 8782 US (Tacoma)\u003Cbr \/\u003E\r\nMeeting ID: 899 6781 4013\u003Cbr \/\u003E\r\nPasscode: 374057\u003Cbr \/\u003E\r\nFind your local number: \u003Ca href=\u0022https:\/\/us02web.zoom.us\/u\/kcYGKDrcOl\u0022\u003Ehttps:\/\/us02web.zoom.us\/u\/kcYGKDrcOl\u003C\/a\u003E\u003C\/p\u003E\r\n\r\n\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":"","field_summary_sentence":[{"value":"Interactive Scalable Interfaces for Machine Learning Interpretability "}],"uid":"27707","created_gmt":"2020-09-24 16:16:41","changed_gmt":"2020-09-24 16:16:41","author":"Tatianna Richardson","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2020-10-05T14:30:00-04:00","event_time_end":"2020-10-05T16:30:00-04:00","event_time_end_last":"2020-10-05T16:30:00-04:00","gmt_time_start":"2020-10-05 18:30:00","gmt_time_end":"2020-10-05 20:30:00","gmt_time_end_last":"2020-10-05 20:30:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"groups":[{"id":"221981","name":"Graduate Studies"}],"categories":[],"keywords":[{"id":"100811","name":"Phd Defense"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78761","name":"Faculty\/Staff"},{"id":"78771","name":"Public"},{"id":"174045","name":"Graduate students"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}