NIPS 2005 Workshop
Large Scale Kernel Machines


Friday December 9th, 2005

The workshop has attracted a large number of high quality submissions. Rather than limiting the number of presentations, we have chosen to organize a poster session. The choice ``talk or poster''  was partly constrained by prior engagements and is not meant to suggest any difference in quality.

Morning session: 7:30am--10:30am

7:30am Welcome
7:40am Working Set Selection Using the Second Order Information for SVMs
Chih-Jen Lin
8:00am Large-scale parallel implementations of SVMs
Hans-Peter Graf, Igor Durdanovic, Eric Cosatto
8:20am Online Learning with a Memory Harness
Ofer Dekel, Shai Shalev-Schwartz, Yoram Singer
8:40am Poster spotlights
9:00am Coffee break and Poster setup
9:10am Implementing SVM in an RDBMS: Improved Scalability and Usability
Joseph S. Yarmus, Boriana L. Milenova, Marcos M. Campos
9:30am Large Scale Genomic Sequence Support Vector Machines
Soren Sonnenburg, Gunnar Rätsch
9:50am Improved Fast Gauss Transform
Ramani Duraiswami, Vikas Raykar
10:10am Approximate Methods for Gaussian Process Regression
Chris Williams, Joaquin Quiñonero Candela, Carl Rasmussen


Afternoon session: 3:30pm--6:30pm

3:30pm Trading Convexity for Scalability
Ronan Collobert, Jason Weston
3:50pm Fast greedy algorithms for building sparse kernel machines
Sathiya Keerthi
4:10pm Online SVM Learning with SMD Gain Adaptation
Nic Schraudolph, S.V.N. Vishwanathan, Alex Smola
4:30pm Coffee Break and Posters:

Simple and SimplerSVM
S. V. N. Vishwanathan
A parallel training algorithm for large scale support vector machines
Elad Tom-Yov
Fast Transpose Methods for Sparse Kernel Learning
Patrick Haffner
"Dual consistency" for training methods in kernel methods
Gökhan Bakır
On-line Learning with Sparse Kernels for Video Analysis
Li Cheng, Shaojun Wang, Dale Schuurmans, Terry Caelli
Very Large Scale Manifold Regularization using Core Vector Machines
Ivor W. Tsang, James T. Kwok
Kernel ICA for Large Scale Problems
Stephanie Jegelka, Arthur Gretton, Dimitris Achlioptas
Distributed objects applied to Support Vector Machines
Joannes Vermorel
5:20pm The Curse of Learning Complicated Functions using Local Kernels
Yoshua Bengio
5:40pm LASVM applied to large invariant problems
Gaelle Loosli
6:00pm Learning large-scale tasks with complicated invariances: Why  kernel methods fall short
Yann LeCun
6:20pm Discussion and Concluding remarks