Tatsuya Matsushima comments

Results 59 comments of


                                            Tatsuya Matsushima

Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagation

## 次に読む Early Inference in Energy-Based Models Approximates Back-Propagation - https://arxiv.org/abs/1510.02777 STDP as Presynaptic Activity Times Rate of Change of Postsynaptic Activity Approximates Backpropagation - STDP-Compatible Approximation of Backpropagation in...

Deep Mean Field Games for Learning Optimal Behavior Policy of Large Populations

論文紹介 - https://hub.packtpub.com/paper-two-minutes-using-mean-field-games-learning-behavior-policy-large-populations/

Predicting Deeper into the Future of Semantic Segmentation

github - https://github.com/facebookresearch/SegmPred

Time-Contrastive Networks: Self-Supervised Learning from Video

プロジェクトページ - https://sermanet.github.io/imitate/ github - https://github.com/tensorflow/models/tree/master/research/tcn youtube - https://www.youtube.com/watch?v=b1UTUQpxPSY スライド - https://docs.google.com/presentation/d/1EvWSbsFfnceBpN7yG1wnqM2LxySQ0Gi-wTx6QaoVekY/edit#slide=id.p

Embodied Question Answering

サイト - https://embodiedqa.org/ シミュレータのcode(github) - https://github.com/facebookresearch/house3d

End-to-end deep image reconstruction from human brain activity

#186 を改良したモデル

Adaptive Path-Integral Autoencoder: Representation Learning and Planning for Dynamical Systems

確率微分方程式(SDE)使ってる

Overcoming catastrophic forgetting in neural networks

DeepMind blog - https://deepmind.com/blog/enabling-continual-learning-in-neural-networks/

VIME: Variational Information Maximizing Exploration

github - https://github.com/openai/vime blog - http://mabonki0725.hatenablog.com/entry/2018/03/04/095146 - `本論文は逆強化学習であることがわかる`

A sensorimotor account of vision and visual consciousness

sensorimotor contingency(SMC)を提案した論文 1.3節 > The central idea of our new approach is that vision is a mode of exploration of the world that is mediated by knowledge of what we...