Programme
Monday, 16 March
10:30 - 10:40 Opening
Chair: Nihat Ay
10:40 - 11:40 (invited talk) Information Geometry and Wasserstein Distance
Shun-ichi Amari, RIKEN Center for Brain Science
11:40 - 13:10 Lunch
Chair: Asuka Takatsu
13:10 - 14:10 (invited talk) Logarithmic divergences: from finance to optimal transport and information geometry
Ting-Kam Leonard Wong, University of Toronto
14:20 - 15:20 (invited talk) On the Natural Gradient for Deep Learning
Nihat Ay, Max Plank Institute for Mathematics in the Sciences
Chair: Ting-Kam Leonard Wong
15:50 - 16:50 (invited talk) On the Berezin-Wallach-Gindikin-Jorgensen set
Hideyuki Ishi, Osaka City University
17:00 - 18:00 (invited talk) T.B.A.
Hiroshi Nagaoka, The University of Electro-Communications
Tuesday, 17 March
Chair: Takashi Takenouchi
10:00 - 11:00 (invited talk) Learning with Dually Flat Structure and Incidence Algebra
Mahito Sugiyama, National Institute of Informatics
11:05 - 11:20 (short talk) Selection bias may be adjusted when sample size is negative
Hidetoshi Shimodaira, Kyoto University
11:20 - 11:35 (short talk) Riemannian Geometry Machine Learning Methods for EEG and fNIRS Digital Biomarker Development in AI for Aging Societies Application
Tomasz Rutkowski, RIKEN AIP
11:40 - 13:10 Lunch
Chair: Takafumi Kanamori
13:10 - 14:10 (invited talk) From geometric learning machines to the geometry of learning
Frank Nielsen, Sony Computer Science Laboratories Inc.
14:20 - 15:20 (invited talk) Multilabel retrieval: a loss function perspective
Aditya Menon, Google Research
15:50 - 16:50 (invited talk) T.B.D.
Sebastian Nowozin, Google Research
17:00 - 18:00 (invited talk) Information Geometry and Optimal Transport in Reproducing Kernel Hilbert Spaces
Minh Ha Quang, RIKEN AIP
18:10 - Bauquet (Registration has been closed)
Wednesday, 18 March
Chair: Jun Zhang
10:00 - 11:00 (invited talk) Information geometry of reinforcement learning
Shinto Eguchi, The Institute of Statistical Mathematics
11:05 - 11:20 (short talk) MDL Estimators Using Fiber Bundles of Local Exponential Families
Junichi Takeuchi, Kyushu University
11:20 - 11:35 (short talk) Non-informative prior on the Kähler information statistical manifold of complex-valued Gaussian process
Hidemasa Oda, The University of Tokyo
11:40 - 13:10 Lunch
Chair: Kei Kobayashi
13:10 - 13:25 (short talk) A Generalization Bound for Online Variational Inference
Pierre Alquier, RIKEN AIP
13:25- 13:40 (short talk) Deformed q-ELBO for Robust Variational Inference
Deepika Kumari, Romanian Institute of Science and Technology
13:40 - 13:55 (short talk) On the gradient-flow equations in information geometry
Tatsuaki Wada, Ibaraki University
13:55 - 14:10 (short talk) On the Natural Gradient for the Training of Neural Networks
Luigi Malagò, Romanian Institute of Science and Technology
Chair: Shinto Eguchi
14:20 - 15:20 (invited talk) Information geometry of multivariate Poisson distributions and its applications
Fumiyasu Komaki, The University of Tokyo
Chair: Kei Kobayashi
15:50 - 16:50 (invited talk) Bayesian Learning Rule: Combining information geometry, optimization, and statistics to improve deep learning
Mohammad Emtiyaz Khan, RIKEN AIP
Chair: Fumiyasu Komaki
17:00 - 18:00 (invited talk) Statistical Mirror Symmetry
Jun Zhang, University of Michigan
Thursday, 19 March
Chair: Milan Mosonyi
10:00 - 10:30 (tutorial) Introduction to Quantum Information
Fuyuhiko Tanaka, Osaka University
10:40 - 11:40 (invited talk) Recent progress in asymptotic quantum statistics
Akio Fujiwara, Osaka University
11:40 - 13:10 Lunch
Chair: Fuyuhiko Tanaka
13:10 - 13:25 (short talk) Natural Alpha Embeddings and their Impact on Downstream Tasks
Riccardo Volpi, Romanian Institute of Science and Technology
13:25 - 13:40 (short talk) Bregman-Lagrangian Dynamics on the Non-parametric Statistical Bundle
Goffredo Chirco, Romanian Institute of Science and Technology
13:40 - 13:55 (short talk) Towards a canonical divergence within Information Geometry
Domenico Felice, MPI MiS - Leipzig
13:55 - 14:10 (short talk) A method to construct exponential families by representation theory
Koichi Tojo, RIKEN AIP
Chair: Akio Fujiwara
14:20 - 15:20 (invited talk) Rényi divergence radii in quantum information theory
Milan Mosonyi, Budapest University of Technology and Economics
Chair: Takeru Matsuda
15:50 - 16:50 (invited talk) Qubit channel parameter estimation with noisy initial states
David Collins, Colorado Mesa University
Chair: David Collins
17:00 - 18:00 (invited talk) Minimax estimation of quantum states based on the conditional Holevo mutual information
Takeru Matsuda, The University of Tokyo
Friday, 20 March
Chair: TBA
10:00 - 11:00 (invited talk) Geometry of Regularized Wasserstein Distances
Marco Cuturi, Google
11:10 - 12:10 (invited talk) Invariant metric under deformed Markov embeddings with overlapped supports
Asuka Takatsu, Tokyo Metropolitan University
12:10 - 12:20 Closing
Poster Presentation (Monday - Friday)
1. Multi-scale k-nearest neighbour
Akifumi Okuno, RIKEN AIP
2. Natural Wake-Sleep Algorithm for Helmholtz Machines
Csongor-Huba Varady, Romanian Institute of Science and Technology, MPI MiS
3. Correlated Uncertainties for Regression Problems using Bayesian Neural Networks and Generalized Divergences
Hector Hortua, Romanian Institute of Science and Technology
4. Information Geometric Regularizers for Variational AutoEncoders to Improve Robustness Against Adversarial Examples
Petru Hlihor, Romanian Institute of Science and Technology and MPI MiS
5. Projection of Bayesian predictive densities onto finite-dimensional exponential families
Michiko Okudo, The University of Tokyo
6. Information geometry of operator scaling
Tasuku Soma, The University of Tokyo