Schedule
NIPS 2015 CiML workshop [new: we have posted the slides below]
Saturday, December 12, 2015
Palais des Congrès de Montréal
Convention and Exhibition Center
Room 512 e [Floor map] Talk abstracts
Morning session (9:00 am-12:00 pm)
9:00 - Welcome and introduction. Evelyne Viegas [slides]
9:10 - Invited talk, Challenges in Medical Image Analysis: Comparison, Competition, Collaboration, Bram van Ginneken [slides]
9:50 - Break
10:20 - Invited talk, Techniques and Technologies for Efficient and Realistic Benchmarks: Examples from the MediaEval Multimedia Benchmark and CLEF NewsREEL, Martha Larson [slides]
11:00 - Discussion: Open Innovation, Balazs Kegl and Ben Hamner moderators [Balazs' post][Balazs' slides]
12:00 - Break
Break-out session on AutoML challenge.
12:30 - Presentation of the AutoML challenge. Isabelle Guyon -- Announcement of the new GPU track. [slides]
13:00 - Automated Machine Learning: Successes & Challenges. Frank Hutter. Team aaad_freiburg. First place AutoML1 phase, second place AutoML2 phase. [paper][supplementary material][slides]
13:30 - Sensible allocation of computation for ensemble construction. James Lloyd. Team jrl44/backstreet.bayes. First place AutoML2 phase, second place AutoML1 phase. [slides]
14:00 - Scalable ensemble learning with stochastic feature boosting. Eugene Tuv. Team ideal.intel.analytics. First place Final0 phase, second place Final1 phase. [slides]
14:30 - Break
Afternoon session (15:00-18:30)
15:00 - Invited talk, Lessons Learned from the PASCAL VOC Challenges, and Improving the Data Analytics Process, Chris Williams [slides]
15:40 - Discussion: Coopetitions, Evelyne Viegas and Isabelle Guyon moderator
16:40 - Break
17:00 - Contributed talk, Academic Torrents: Scalable Data Distribution, Henry Z. Lo and Joseph Paul Cohen [paper][slides]
17:30 - Open discussion, Michele Sebag modelator
18:20 - Wrap up
18:30 - Adjourn
Eugene Tuv
James Lloyd
Chris Willians
Balazs Kegl
Joseph P. Cohen and Henry Z. Lo
Ben Hamner
Frank Hutter
Bram van Ginneken
Erick Watson
Joseph, Henry & Sebastien Treger
Jaffray Woodriff & Bram
Balazs, John Platt & Chris
We are connected to the Bayesian Optimization workshop and the Black Box Learning and Inference workshop, because they both treat in some way the "Automatic Machine Learning" problem, which we will discuss during the lunch session.
We are grateful to our Committee for helping put this program together.