THANK YOU to all who contributed to the success of eLUCID8!
Data Science and Human Behavior | In the Lab and In the Wild
Wisconsin Institute for Discovery, Madison, WI
August 14th – August 15th, 2017
eLUCID8 featured interactive presentations, talks, and roundtables intended to discuss LUCID related projects and potential collaborations with government agencies, non-profits and industry groups.
Scroll down or click for the Agenda.
For communication during the conference (your questions, announcements and evaluation links) please sign up for eLUCID on Slack: https://join.slack.com/t/elucid8/signup
We truly appreciate and value your feedback. Please let us know what your experience was like at eLUCID8 in the following short surveys: Monday eLUCID8, Tuesday eLUCID8. Our students also respect and honor your feedback for their presentations. Please let us know what you think: LUCID Presentations
Monday, August 14th
Former Cartoon Editor of The New Yorker,
Present Cartoon and Humor Editor of Esquire
Humor is traditionally at the hands of its author. What happens when the audience picks the punchline?
Each week, on the last page of the magazine, The New Yorker provides a cartoon in need of a caption. Readers submit captions, the magazine chooses three finalists, readers vote for their favorites. It’s humor—crowdsourced—and with more than 3 million submissions provided by 600,000 participants, it provides tremendous insight as to what makes us laugh.
In a fast-paced and funny talk, Bob Mankoff, The New Yorker‘s cartoon editor, will analyze the lessons we learn from crowdsourced humor. Along the way, he’ll explore how cartoons work (and sometimes don’t); how he makes decisions about what cartoons to include; and what crowds can tell us about a good joke.
“Amplifying Human Capabilities on Visual Categorization Tasks”
We are developing methods to improve human learning and performance on challenging visual categorization tasks, e.g., bird species identification, diagnostic dermatology. Our approach involves inferring _psychological embeddings_ — internal representations that individuals use to reason about a domain. Using predictive cognitive models that operate on an embedding, we perform surrogate-based optimization to determine efficient and effective means of training domain novices as well as amplifying an individual’s capabilities at any stage of training. Our cognitive models leverage psychological theories of: similarity judgement and generalization, contextual and sequential effects in choice, attention shifts among embedding dimensions. Rather than searching over all possible training policies, we focus our search on policy spaces motivated by the training literature, including manipulation of exemplar difficulty and the sequencing of category labels. We show that our models predict human behavior not only in the aggregate but at the level of individual learners and individual exemplars, and preliminary experiments show the benefits of surrogate-based optimization on learning and performance.
This work was performed in collaboration with Brett Roads at the University of Colorado.
Michael Mozer received a Ph.D. in Cognitive Science at the University of California at San Diego in 1987. Following a postdoctoral fellowship with Geoffrey Hinton at the University of Toronto, he joined the faculty at the University of Colorado at Boulder and is presently an Professor in the Department of Computer Science and the Institute of Cognitive Science. He is secretary of the Neural Information Processing Systems Foundation and has served as chair of the Cognitive Science Society. He is interested both in developing machine learning algorithms that leverage insights from human cognition, and in developing software tools to optimize human performance using machine learning methods.
Monday August 14th
9:00-9:30 Welcome to eLUCID8 from our LUCID Director, Tim Rogers
9:30-10:30 Learning in Childhood with Jenny Saffran, Ed Hubbard and Chuck Kalish
10:45-Noon Machine Learning & Human Behavior with Varun Jog, Dimitris Papailiopoulos, Joe Austerweil and Jerry Zhu
1:30-2:15 Making Sense of the Ineffable with Karen Schloss and Paula Niedenthal
2:15-2:45 Data Science in the Wild with our Keynote Speakers Bob Mankoff and Michael Mozer
3:00-4:15 Data Blitz
4:15-5:00 Poster and Movie Session
5:00-6:00 Posters, Mingling and Cash Bar
6:00-7:00 KEYNOTE: Bob Mankoff
Tuesday August 15th
9:30-10:30 Science Communication Panel with Veronica Reuckert, host of “Central Time” on WPR, Jordan Ellenberg, author of How Not to be Wrong: The Power of Mathematical Thinking, and Mark Seidenberg, author of Language at the Speed of Sight: How We Read, Why So Many Can’t, and What Can Be Done About It
10:30-12:00 LUCID Science Talks
1:00-1:45 University-Industry Partnerships from Susan LaBelle, UW–Madison Office of Corporate Relations
2:00-3:20 Data Science in the Wild with City of Madison, 4W, POLCO, and Lands’ End
3:20-4:00 Breakouts and Group Discussion
4:00-5:00 KEYNOTE: Michael C. Mozer