ICML 2010 - Awards

Every year, ICML honors its best contributions with best paper awards. In 2010, the selection committee chose the contributions listed below. All best paper award winners received a certificate and a check for $1000, and all runner-ups received a check for $500. We are grateful to Springer, the publishers of Machine Learning, for helping to fund these awards.

Best Paper Award

For the Best Paper Award, we prepared a list of 19 papers that received excellent reviews from the reviewers and the area chair. In a two-step selection process, the following paper emerged as the clear winner of the ICML 2010 Best Paper Award:

• Le Song, Byron Boots, Sajid Siddiqi, Geoffrey Gordon, Alex Smola:
  Hilbert Space Embeddings of Hidden Markov Models
  

The runner-up in this category was the following paper:

• Julia Vogt, Sandhya Prabhakaran, Thomas Fuchs, Volker Roth:
  The Translation-invariantWishart-Dirichlet Process for Clustering Distance Data
  

Best Student Paper Award

The best student papers were selected from the same 19 papers as the best paper awards, again in a two-step selection process. Only those papers that had a student as the primary author were eligible. The following paper won the best student paper award:

• John Duchi, Lester Mackey, Michael Jordan:
  On the Consistency of Ranking Algorithms
  

The close runner-up was:

• Brian Ziebart, Drew Bagnell, Anind Dey:
  Modeling Interaction via the Principle of Maximum Causal Entropy
  

Best Application Paper Award

For the best application, we formed a short list of six candidates with a strong focus on applications. We were excited to give the award to a paper that brings together basic research in transfer learning and cognitive science:

• Kevin Canini, Mikhail Shashkov, Tom Griffiths:
  Modeling Transfer Learning in Human Categorization with the Hierarchical Dirichlet Process
  

The runner-up in this category was:

• Ali Shoeb, John Guttag:
  Application of Machine Learning to Epileptic Seizure Detection
  

Best 10-Year Paper

Last but not least, we had the privilege of honoring the most influential paper of ICML 2000 (Stanford). In a first phase, we selected a short list of six papers based on suggestions from the best paper committee and several senior members of the community. We then invited all area chairs of ICML 2010 and all members of the IMLS board to name and rank their three favorite papers among these six. From this vote, the following paper emerged as a clear winner:

• Erin L. Allwein, Robert E. Schapire, Yoram Singer:
  Reducing Multiclass to Binary: A Unifying Approach for Margin Classifiers

We were glad to have Rob Schapire take a look back upon this great piece of work and discuss its influence on the community.