They present a new paper: “PAC-Bayesian Theory for Transductive Learning“, co-authored with Luc Bégin and François Laviolette. Additionally, they will present two posters on Risk Bounds for the Majority Vote on Thursday’s joint AISTATS/MLSS poster session.
The GRAAL will be at the 2013 edition of Journée Phare in Orford, Quebec, Canada. Sébastien Giguère will give an talk on his recent work on high bioactivity peptide design and screening and Alexandre Drouin will present a poster on MHC-NP, a tool for predicting peptides naturally processed by the major histocompatibility complex. We hope […]
The immune epitope database (IEDB), the largest public database in the field of immunology, now offers MHC-NP as part of its online analysis tools. Users can upload protein sequences in fasta format to query the predictor to find out which peptides are most likely to be naturally processed by the Type I Major Histocompatibility Complex […]
The GRAAL members will present two papers at ICML 2013 conference (Atlanta, 16-21 june): A PAC-Bayesian Approach for Domain Adaptation with Specialization to Linear Classifiers Authors: Pascal Germain, Amaury Habrard, François Laviolette, Emilie Morvant Talk: Tuesday, June 18, Track C (Transfer Learning, 4:00 to 5:40) Risk Bounds and Learning Algorithms for the Regression Approach to […]