OPT 2010: 3rd International Workshop on Optimization for Machine Learning

Time	Speaker	Title	Slides	Video
7:30-7:40	Organizers	Opening remarks
07:40--08:30	Mark Schmidt	Invited talk: Limited-memory quasi-Newton and Hessian-free Newton methods for non-smooth optimization.	[Slides]
08:30--08:50	Alekh Agarwal	Information-theoretic lower bounds on the oracle complexity of sparse convex optimization	[Slides]
09:00--09:30	COFFEE BREAK
09:30--10:20	Yurii Nesterov	Invited Talk:Efficiency of Quasi-Newton Methods on Strictly Positive Functions.	[Slides]
10:20--10:30	Poster Spotlights
	Yunlong He	An efficient algorithm for sparse PCA
	Matthieu Geist	Statistical Linearization for Value Function Approximation in Reinforcement Learning
	Gregory Moore	Gradient-type methods for primal SVM model selection
10:30--15:30	Long Break; Poster Session Begins
15:30--16:20	Laurent El Ghaoui	Invited Talk: Safe Feature Elimination in Sparse Learning
16:20--16:40	Dengyong Zhou	Hierarchical Classification via Orthogonal Transfer	[Slides]
16:40--17:00	Shayok Chakraborty	An Optimization Based Framework for Dynamic Batch Mode Active Learning.pdf	[Slides]
17:00--17:30	COFFEE BREAK
17:30--17:50	Andre Martins	Augmenting Dual Decomposition for MAP Inference.	[Slides]
17:50--18:10	Jeremy Jancsary	An Incremental Subgradient Algorithm for Approximate MAP Estimation in Graphical Models.	[Slides]
18:10--20:00	Poster Session Continues

Updates

Please don't forget to Register
Workshop Location: TBA
Workshop Date: 10.dec.2010

Poster Spotlights

• An efficient algorithm for sparse PCA
• Statistical Linearization for Value Function Approximation in Reinforcement Learning
• Gradient-type methods for primal SVM model selection