Schedule

October 11, 2018 - BayLearn 2018 Symposium

8:00am

Registration opens

8:00am - 9:00am

Coffee & Light Breakfast

9:05am  

Welcome to BayLearn 2018! 

9:15am

Keynote 1


Data-Driven Dialogue Systems: Models, Algorithms, Evaluation, and Ethical Challenges
Joelle Pineau, Facebook & McGill University

10:00am  

Chair: Jerremy Holland

Session 1



Doubly Reparameterized Gradient Estimators for Monte Carlo Objectives
George Tucker, Dieterich Lawson, Chris J Maddison 
Local Linear Forests: Leveraging Smoothness with Random Forests
Rina S Friedberg, Julie Tibshirani, Susan Athey, Stefan Wager
Deep Learning for Supercomputers
Noam M Shazeer, Youlong Cheng, Niki Parmar, Dustin Tran, Ashish Vaswani, Penporn Koanantakool, Peter Hawkins, HyoukJoong Lee, Mingsheng Hong, Cliff Young, Ryan Sepassi, Blake Hechtman 

10:45am - 11:10am

Break

11:15am  

Keynote 2


Some representation, optimization and generalization properties of deep networks
Peter Bartlett, CS and Statistics, UC Berkeley

12:00pm  

Chair: Jean-Francois Paiement

Session 2


A Deep Generative Model for Semi-Supervised Classification with Noisy Labels
Maxime Langevin
Improved Mixed-Example Data Augmentation
Cecilia Summers

12:30pm - 1:50pm

Lunch

2:00pm

Keynote 3


Statistical and machine learning challenges from genetics to CRISPR gene editing
Jennifer Listgarten, Center for Computational Biology & Berkeley AI Research, UC Berkeley

2:45pm

Chair: Isabelle Guyon

Session 3

 


Plannable Representations with Causal InfoGAN
Thanard Kurutach, Aviv Tamar
Deep Reinforcement Learning in a Handful of Trials using Probabilistic Dynamics Models
Kurtland Chua, Roberto Calandra, Rowan McAllister, Sergey Levine 

3:15pm - 3:40pm

Break

3:45pm    

Chair: David Grangier

Keynote 4


What's Wrong with Meta-Learning (and how we can fix it)
Sergey Levine, Berkeley AI Research Lab, UC Berkeley and Google

4:45pm

Main Poster Session, Prizes, Food and Refreshments

7:00pm

End of the Symposium