2019/7/23 16:08
要旨
Abstract
The purpose of this workshop is to bring together discrete optimizers and machine learning researchers, theory people and practitioners in an effort to bridge the fields and stimulate cross-disciplinary interaction.
Program
Monday, July 29
| 09h30 – 10h00 | Reception/ Opening remarks | |
| 10h00 – 12h00 | Session A1 Chair: Amitabh Basu | |
| A11 | An approximation algorithm for training one-node ReLU neural network/ Santanu Dey (Georgia Tech) | |
| A12 | Subspace communication driven search for high dimensional optimization/ Logan Mathesen (Arizona State University) | |
| A13 | A limiting analysis of regularization of SDP and its implication to infeasibleinterior-point algorithms/ Takashi Tsuchiya (Graduate Research Institute for Policy Studies) | |
| A14 | Stochastic proximal methods for non-smooth non-convex constrained sparseoptimization/ Michael Metel (RIKEN AIP) | |
| 12h00 – 13h00 | Lunch | |
| 13h00 – 14h30 | Session A2 Chair: Santanu Dey | |
| A21 | Locally accelerated conditional gradients/ Alejandro Carderera (Georgia Tech) | |
| A22 | No-regret algorithms for onlinek-submodular maximization/ Tasuku Soma (University of Tokyo) | |
| A23 | Parallel depth-first search and the applications to data mining/ Kazuki Yoshizoe (RIKEN AIP) | |
| 14h30 – 15h00 | Coffee break | |
| 15h00 – 16h30 | Session A3 Chair: Giulia Pedrielli | |
| A31 | Bayesian deep learning in portfolio optimization/ Akshay Gupte (Clemson University) | |
| A32 | Learning-algorithms from Bayesian principles/ Emtiyaz Khan (RIKEN AIP) | |
| A33 | Matrix co-completion for multi-label classification with missing featuresand labels/ Miao Xu (RIKEN AIP) | |
Tuesday, July 30
| 09h30 – 10h00 | Reception | |
| 10h00 – 12h00 | Session B1 Chair: Masashi Sugiyama | |
| B11 | Statistical decision theory perspectives on learning and stochastic optimiza-tion/ Amitabh Basu (Johns Hopkins University) | |
| B12 | On solving mixed integer non linear programs with separable non convexi-ties/ Claudia D’Ambrosio (CNRS & Ecole Polytechnique) | |
| B13 | Blended matching pursuit/ Cyrille Combettes (Georgia Tech) | |
| B14 | Simulation based mixed integer programming/ Alexander Martin (Universit ̈at Erlangen-N ̈urnberg) | |
| 12h00 – 13h00 | Lunch | |
| 13h00 – 14h30 | Session B2 Chair: Edwin Romeijn | |
| B21 | Strong mixed-integer programming formulations for trained neural net-works/ Joey Huchette (Massachusetts Institute of Technology) | |
| B22 | Airline schedule planning under infrastructure constraint and with customerchoice evaluation/ S ́ebastien Deschamps (Ecole Nationale des Ponts et Chauss ́ees) | |
| B23 | Adding variables – speed up by including new binary variables/ Robert Hildebrand (Virginia Tech) | |
| 14h30 – 15h00 | Coffee | |
| 15h00 – 17h00 | Session B3 Chair: Alexander Martin | |
| B31 | Causal-retro-causal Systems induce a dynamics on a manifold/ Hans-Georg Zimmermann (Fraunhofer N ̈urnberg) | |
| B32 | Adaptive algorithm for finding connected dominating sets in uncertaingraphs/ Takuro Fukunaga (Chuo University) | |
| B33 | Computing full conformal prediction set with approximate homotopy/ Eug`ene Ndiaye (RIKEN AIP) | |
| B34 | Differentiable ranks using optimal transport: The Sinkhorn CDF and quan-tile operator/ Marco Cuturi (Google Brain) | |
Wednesday, July 31
| 09h30 – 10h00 | Reception | |
| 10h00 – 12h00 | Session C1 Chair: Claudia D’Ambrosio | |
| C11 | Sampling fromlogsupermodular distribution/ Shuji Kijima (Kyushu University) | |
| C12 | Fair dimensionality reduction and iterative rounding for SDPs/ Uthaipon Tantipongpipat (Georgia Tech) | |
| C13 | Random projections for quadratic programs/ Leo Liberti (CNRS & Ecole Polytechnique) | |
| C14 | Random projection for conic programming/ Pierre-Louis Poirion (RIKEN AIP) | |
| 12h00 – 14h00 | Lunch | |
| 14h00 – 15h30 | Session C2 Chair: Leo Liberti | |
| C21 | Multi-armed bandit problem in piece-wise stationary environment/ Chi-Guhn Lee (University of Toronto) | |
| C22 | Policy transfer via greedy state recoding/ Christopher Mutschler (Universit ̈at Erlangen-N ̈urnberg) | |
| C23 | Stochastic monotone submodular maximization with queries/ Takanori Maehara (RIKEN AIP) | |
| 15h30 – 16h00 | Coffee | |
| 16h00 – 17h30 | Session C3 Chair: Joey Huchette | |
| C31 | Tensor network representation in machine learning/ Qibin Zhao (RIKEN AIP) | |
| C32 | Machine learning approaches in brain correlates of dementia elucidation -tensor machine lLearning and beyond/ Tomasz Rutkowski (University of Tokyo / RIKEN AIP) | |
| C33 | Towards robust deep learning/ Bo Han (RIKEN AIP) | |
| 17h30 | Closing remarks | |
(Last update: July 23, 2019)
詳細情報
| 日時 | 2019/07/29(月) 10:00 - 16:30 2019/07/30(火) 10:00 - 16:30 2019/07/31(水) 10:00 - 17:30 |
| URL | http://www.cas.mcmaster.ca/~deza/tokyo2019.html |
