Travelled to:1 × France
1 × Germany
1 × USA
1 × United Kingdom
2 × China
Collaborated with:T.S.Jaakkola R.Urtasun A.Cohen A.G.Schwing M.Pollefeys A.Shashua M.Osadchy D.Keren F.Orabona A.D.Sarwate J.Peng D.A.McAllester
Talks about:perturb (3) posteriori (2) structur (2) maximum (2) random (2) hyperplan (1) algorithm (1) parallel (1) function (1) concentr (1)
Person: Tamir Hazan
DBLP: Hazan:Tamir
Contributed to:
201520142014201220112005Wrote 8 papers:
- ICML-2015-CohenH #learning #online
- Following the Perturbed Leader for Online Structured Learning (AC, TH), pp. 1034–1042.
- ICML-2015-OsadchyHK #classification
- K-hyperplane Hinge-Minimax Classifier (MO, TH, DK), pp. 1558–1566.
- ICML-c1-2014-OrabonaHSJ #on the #random
- On Measure Concentration of Random Maximum A-Posteriori Perturbations (FO, TH, ADS, TSJ), pp. 432–440.
- ICML-c2-2014-SchwingHPU #algorithm #convergence #parallel #using
- Globally Convergent Parallel MAP LP Relaxation Solver using the Frank-Wolfe Algorithm (AGS, TH, MP, RU), pp. 487–495.
- ICML-2012-HazanJ #on the #random
- On the Partition Function and Random Maximum A-Posteriori Perturbations (TH, TSJ), p. 217.
- ICML-2012-SchwingHPU #modelling #performance #predict #visual notation
- Efficient Structured Prediction with Latent Variables for General Graphical Models (AGS, TH, MP, RU), p. 216.
- ICML-2011-PengHMU #set
- Convex Max-Product over Compact Sets for Protein Folding (JP, TH, DAM, RU), pp. 729–736.
- ICML-2005-ShashuaH #statistics
- Non-negative tensor factorization with applications to statistics and computer vision (AS, TH), pp. 792–799.