SIERRA - 2013 - Annual activity report

SIERRA

SIERRA - 2013

Project-Team Sierra

Members

Overall Objectives

Statement
Highlights of the Year

Research Program

Supervised Learning
Unsupervised Learning
Parsimony
Optimization

Application Domains

Application Domains
Natural Language Processing

Software and Platforms

SPAMS (SPArse Modeling Software)
BCFWstruct
SAG
fMRI

New Results

Block-Coordinate Frank-Wolfe Optimization for Structural SVMs
Minimizing Finite Sums with the Stochastic Average Gradient.
Fast Convergence of Stochastic Gradient Descent under a Strong Growth Condition
Non-strongly-convex smooth stochastic approximation with convergence rate $O (1 / n)$
Streaming Bayesian Inference
Convex Relaxations for Permutation Problems
Phase retrieval for imaging problems
Learning Sparse Penalties for Change-point Detection using Max Margin Interval Regression
Maximizing submodular functions using probabilistic graphical models
Reflection methods for user-friendly submodular optimization
Convex Relaxations for Learning Bounded Treewidth Decomposable Graphs
Large-Margin Metric Learning for Partitioning Problems
Comparison between multi-task and single-task oracle risks in kernel ridge regression
Sharp analysis of low-rank kernel matrix approximations
fMRI encoding and decoding models
Structured Penalties for Log-linear Language Models
Distributed Large-scale Natural Graph Factorization
Evaluating Speech Features with the Minimal-Pair ABX task
Hidden Markov Tree Models for Semantic Class Induction
Domain Adaptation for Sequence Labeling using Hidden Markov Models
Simple Greedy Matching for Aligning Large Knowledge Bases

Bilateral Contracts and Grants with Industry

Bilateral Contracts with Industry
Bilateral Grants with Industry

Partnerships and Cooperations

National Initiatives
European Initiatives
International Initiatives
International Research Visitors

Dissemination

Scientific Animation
Teaching - Supervision - Juries
Popularization

Bibliography

Publications of the year
References in notes

Previous |

Home | Next next

Section: Software and Platforms

BCFWstruct

Participants : Simon Lacoste-Julien [correspondent] , Mark Schmidt.

BCFWstruct is a Matlab implementation of the Block-Coordinate Frank-Wolfe solver for Structural SVMs. See the ICML 2013 paper with the same name.

Participants outside of Sierra: Martin Jaggi (Centre de Mathématiques Appliquées, Ecole Polytechnique); Patrick Pletscher (Machine Learning Laboratory, ETH Zurich)

Previous |

Home | Next next