Schedule

Session chair: Hugo Larochelle
  • 9:00-9:20 Welcome
  • 9:20-9:50 Invited Talk: Holger Hoos --- Programming by Optimisation: Automated Design and Customisation of Learning Procedures
  • 9:50-10:20 Invited Talk: Jasper Snoek --- New Methods in Bayesian Optimization for Machine Learning
10:20-10:40 Coffee Break

Session chair: Frank Hutter
  • 10:40-11:10 Invited Talk: Vikash Mansinghka --- A View on Automatic Machine Learning from Probabilistic Programming
  • 11:10-11:35 Poster Spotlights 1 (6 spotlights)
    • Automatic Inference for Inverting Software Simulators via Probabilistic Programming
. (Ardavan Saeedi, Vlad Firoiu, Vikash Mansinghka
)
    • Hyperopt-Sklearn: Automatic Hyperparameter Configuration for Scikit-Learn. 
(Brent Komer, James Bergstra, Chris Eliasmith)
    • Sequential Model-Based Ensemble Optimization
. (Alexandre Lacoste, Hugo Larochelle, Mario Marchand, François Laviolette
)
    • Bayesian Optimization with Inequality Constraints
. (Jacob Gardner, Matthew Kusner, Zhixiang Xu, Kilian Weinberger, John Cunningham
)
    • Cognitive Automation of Data Science. (Horst Samulowitz, Chandra Reddy, Ashish Sabharwal
)
    • Preliminary Evaluation of Hyperopt Algorithms on HPOLib. (
James Bergstra, Brent Komer, Chris Eliasmith, David Warde-Farley
)
  • 11:35 -12:00 Poster Session 1
12:00 - 14:00 Lunch

 Session chair: Balázs Kégl
  • 14:00-14:30 Inivted Talk: Dan Roth --- Learning based Programming: Facilitating the Programming of Data Driven Software Systems
  • 14:30-14:55 Poster Spotlights 2 (6 spotlights)
    • Justification Narratives for Individual Classifications. (Or Biran, Kathleen McKeown
)
    • searchspaces: a Python-based canonical description language for hyperparameter search spaces. (David Warde-Farley, James Bergstra)
    • Parameter Inference Engine (PIE) on the Pareto Front. (
Ser Nam Lim, Albert Y. C. Cheng, Xingwei Yang)
    • Universal Unsupervised Transfer Learning through Non-negative Matrix Factorization. (Ievgen Redko, Younès Bennani
)
    • Extrapolating Learning Curves of Deep Neural Networks. (Tobias Domhan, Tobias Springenberg, Frank Hutter
)
    • Automatic Differentiation of Algorithms for Machine Learning. (Atilim Gunes Baydin, Barak Pearlmutter
)
  • 14:55-15:20 Poster Session 2

 15:20-15:40 Coffee Break

 Session chair: Rémi Bardenet

  • 15:40-16:10 Invited Talk: Yoshua Bengio --- Hyper-Parameter Optimization for Deep Learning
  • 16:10-16:40 Announcement of AutoML challenge (Isabelle Guyon)
  • 16:40-17:20 Open Discussion/Panel Discussion (Moderator: Isabelle Guyon)
  • 17:20-17:30 Wrap Up & Announcements about publications, future workshops, …
Comments