The 4th University of Michigan
Workshop on Data Mining
Registration is now closed.
Wednesday, April 24, 2013 (8:30 AM - 5 PM)
Tishman Hall, Bob and Betty Beyster Building, North Campus
1670 CSE building
Sponsored by CSE, Yahoo!, and Office of Research Cyberinfrastructure (ORCI)
Contacts: Mike Cafarella and Drago Radev (send mail to dm2013help, found at umich.edu)
Registration Is Now Closed
Thanks for your interest. We're filled to capacity, so the workshop registration is now closed.
For the last 3 years we have had a big success with the University of
Michigan Workshop on Data Mining. We
uncovered lots of people at Michigan doing interesting and unexpected
things with data. The set of people doing this work is only growing, so
let's get together and again share our work with each other.
Who is invited?
All UM faculty, staff, and graduate students working in the fields of
data mining, broadly construed. We are interested in techniques: models and technologies for
statistical data analysis, Web search technology, analysis of user
behavior, data visualization, etc. We are also interested in
data-centric applications to problems in your own field,
whether it is in the natural sciences, the social sciences, or something else.
External visitors from the State of Michigan and beyond are also welcome to
The workshop will be held at the Computer Science and Enginering Division (CSE) in the Bob and Betty Beyster Building on the North Campus of the University of Michigan. The address is 2260 Hayward Street in Ann Arbor, 48109-2121.
View Larger Map
Parking is available near the Beyster Building but is in short demand because of nearby construction. This map shows available parking that is near the building. Areas marked in blue on the map below require a U-M blue pass or a U-M visitor parking pass. Red areas are metered parking.
Visitor parking passes are available only for those attendees who are external to the University and who requested a pass during registration. If this is you, please park and come into the building and to the registration table in Tishman Hall, the first floor of the atrium in the Beyster Building, in order to pick up your pass.
- April 22, 2013 We now have our workshop agenda
- 8:35 am to 8:45 am - Michael Cafarella, CSE, Intro and Welcome
- 8:45 am to 10:25 am - Technical Talks 1
- Emily Mower Provost, CSE
Using Emotional Noise to Uncloud Audio-Visual Emotion Perceptual Evaluation
- Avishay Livne, CSE and School of Information
Predicting Impact in Academia
- Honglak Lee, CSE
Learning and Selecting Features Jointly with Pointwise Gated Boltzmann Machines
- Perry Samson, AOSS
Mining My Students' Notes to Create Study Guides
- 10:25 am to 10:45 am - Break 1
- 10:45 am to 12:25 pm - Technical Talks 2
- Anna Gilbert, Mathematics
Streaming and Sketching Big Data
- Ashutosh Nandeshwar, Office of University Development
Three Big Problems in Fundraising and How to Solve Them With Data Science
- Zhe Zhao, CSE and School of Information
Questions about Questions: An Empirical Analysis of Information Needs on Twitter
- Brock Palen, ORCI
National Resources for Big Data Exploration
- 12:25 pm to 1:30 pm - Lunch
- 1:30 pm to 3:10 pm - Technical Talks 3
- Jian Tang, School of Information
"Look, Mum, No Hands!" A Parameter-Free Topic Model
- Ambuj Tewari, Statistics
Foundations of Learning to Rank
- Matthew Burgess, CSE
Leveraging Noisy Lists for Social Feed Ranking
- Jason Owen-Smith, Sociology
Data Mining in the Sociology of Science and Organizations
- 3:10 pm to 3:30 pm - Break 2
- 3:30 pm to 4:30 pm - Keynote Address: Scott Gaffney, Director, Knowledge and User Engagement Science, Yahoo! Labs
Title: Personalization at Scale
Media can be loosely defined as the ecosystem that distributes entertainment and information about the world in a timely manner to the mass public. The modern Media ecosystem has been focused on 'mass audiences' for two reasons: 1) it is expensive and time consuming to produce and distribute niche media, and 2) revenue within the ecosystem flows from advertisers/marketers who have been trained to think at the mass audience level. In this talk, we briefly outline the differences between the digital media ecosystem and the traditional media ecosystem, and how this led to a change in how Media was delivered. We then motivate the need to move away from the old view of media audiences to a view of an audience-of-one in which automated personalization becomes a key piece connecting users to the products they use. Finally, we discuss the scale of the personalization problem across Media and detail what Yahoo! is currently doing to deliver true personalization at scale to their users.
- 4:30 pm - Michael Cafarella, CSE, Closing Remarks
- April 11, 2013
Registration is now open!!Registration is now closed.
- April 10, 2013 We have a partial list of speakers. They will include:
- Anna Gilbert, Mathematics, on Streaming and Sketching Big Data
- Brock Palen, ORCI, on National Resources for Big Data Exploration
- Perry Samson, AOSS, on Mining My Students' Notes to Create Study Guides
- Ambuj Tewari, Statistics, on Foundations of Learning to Rank
- Emily Mower Provost, CSE, on Using Emotional Noise to Uncloud Audio-Visual Emotion Perceptual Evaluation
- Honglak Lee, CSE, on Learning and Selecting Features Jointly with Pointwise Gated Boltzmann Machines
- ...and many more!
However, we still have room for a few more speakers. We are particularly interested in talks on biomedical topics. If you are interested in giving a talk, please fill out this short form. Sorry, speaker registration is closed.
- March 12, 2013 The workshop will comprise a series of 25-minute talks, plus a student poster session. We may also be able to have a small number of extended talks for topics that have especially broad appeal.
Our immediate goal is to find a set of interesting speakers. We would like to have speakers from across many different departments on campus. Accordingly, we ask speakers to present a talk that will be interesting to a diverse audience.
If you are interested in giving a talk, please fill out this short form.
If you have presented in the past but have new results you'd like to discuss, please go ahead and fill out the form. Although we expect most presenters will be affiliated with the University of Michigan, we are very interested in hearing from local industry and other universities in the area.