NIPS 2012 Workshop: Perturbations, Optimization, and Statistics

Schedule

7:30-7:40	Danny Tarlow Toronto	Introduction
7:40-8:10	George Papandreou UCLA (invited)	Perturb-and-MAP Random Fields [abstract] [slides]
8:10-8:40	Max Welling UCI and UvA (invited)	Herding versus Perturb-and-MAP [abstract] [slides]
8:40-9:10	Amir Globerson HUJI (invited)	A Simple Geometric Interpretation of SVM using Stochastic Adversaries [abstract]
9:10-9:30	Coffee Break
9:30-10:00	Andrea Montanari Stanford (invited keynote)	Approximate Message Passing Algorithms [abstract] [slides]
10:00-10:30	Tamir Hazan TTI-C (invited)	Learning with Random Maximum A-Posteriori Perturbations [abstract] [slides]
10:30-10:40	Nitish Srivastava Toronto (spotlight)	Improving Neural Networks by Preventing Co-adaptation of Feature Detectors
10:40-10:50	John Duchi Berkeley (spotlight)	Randomized Smoothing for (Parallel) Stochastic Optimization
10:50-12:00	Poster Session (contributed papers)
15:30-16:00	Pascal Vincent Montreal (invited)	Learning Robust Representations Through Input Perturbations: From Stochastic Perturbations to Analytic Criteria and Back [abstract] [slides]
16:00-16:30	Ryan Adams Harvard (invited)	Building Probabilistic Models Around Deterministic Optimization Procedures [abstract] [slides]
16:30-17:00	Shie Mannor Technion (invited)	All Learning is Robust [abstract] [slides]
17:00-17:30	Coffee Break
17:30-18:30	Panel discussion + Wrapup

The following contributed papers will be presented at the poster session.

John C. Duchi, Peter L. Bartlett, Martin J. Wainwright
Randomized Smoothing for (Parallel) Stochastic Optimization (also spotlight)
Geoffrey E. Hinton, Nitish Srivastava, Alex Krizhevsky, Ilya Sutskever, Ruslan R. Salakhutdinov
Improving neural networks by preventing co-adaptation of feature detectors (also spotlight)
Maayan Harel, Shie Mannor
The Perturbed Variation
Alexandre Lacoste, Francois Laviolette, Mario Marchand
Model Averaging With Holdout Estimation of the Posterior Distribution
Ying Liu, Oliver Kosut, Alan S. Willksy
Sampling GMRFs by Subgraph Correction
Varun Ramakrishna, Dhruv Batra
Mode-Marginals: Expressing Uncertainty via Diverse M-Best Solutions
Rajesh Ranganath, Chong Wang, David M. Blei, Eric P. Xing
Adaptively Setting the Learning Rate in Stochastic Variational Inference
Max Siegel, Josh Tenenbaum, Vikash Mansinghka
Improving the Tractability of Bayesian Inverse Graphics by Injecting Noise