no code implementations • 4 Feb 2024 • Koji Hashimoto, Yuji Hirono, Akiyoshi Sannai
Understanding the inner workings of neural networks, including transformers, remains one of the most challenging puzzles in machine learning.
1 code implementation • 2 Feb 2024 • Masayuki Takayama, Tadahisa Okuda, Thong Pham, Tatsuyoshi Ikenoue, Shingo Fukuma, Shohei Shimizu, Akiyoshi Sannai
In practical statistical causal discovery (SCD), embedding domain expert knowledge as constraints into the algorithm is widely accepted as significant for creating consistent meaningful causal models, despite the recognized challenges in systematic acquisition of the background knowledge.
1 code implementation • 31 Jan 2024 • Toshinori Kitamura, Tadashi Kozuno, Masahiro Kato, Yuki Ichihara, Soichiro Nishimori, Akiyoshi Sannai, Sho Sonoda, Wataru Kumagai, Yutaka Matsuo
We study a primal-dual reinforcement learning (RL) algorithm for the online constrained Markov decision processes (CMDP) problem, wherein the agent explores an optimal policy that maximizes return while satisfying constraints.
no code implementations • 21 Sep 2023 • Ryutaro Yamauchi, Sho Sonoda, Akiyoshi Sannai, Wataru Kumagai
In this paper, we propose a novel framework that integrates the Chain-of-Thought (CoT) method with an external tool (Python REPL).
no code implementations • 23 May 2022 • Akiyoshi Sannai, Yasunari Hikima, Ken Kobayashi, Akinori Tanaka, Naoki Hamada
In this paper, we propose a strategy to construct a multi-objective optimization algorithm from a single-objective optimization algorithm by using the B\'ezier simplex model.
no code implementations • 15 Oct 2021 • Akiyoshi Sannai, Makoto Kawano, Wataru Kumagai
We construct learning models based on the reductive Reynolds operator called equivariant and invariant Reynolds networks (ReyNets) and prove that they have universal approximation property.
no code implementations • 29 Sep 2021 • Akiyoshi Sannai, Makoto Kawano, Wataru Kumagai
To overcome this difficulty, we consider representing the Reynolds operator as a sum over a subset instead of a sum over the whole group.
no code implementations • 10 Apr 2021 • Akinori Tanaka, Akiyoshi Sannai, Ken Kobayashi, Naoki Hamada
B\'ezier simplex fitting algorithms have been recently proposed to approximate the Pareto set/front of multi-objective continuous optimization problems.
no code implementations • ICLR 2021 • Makoto Kawano, Wataru Kumagai, Akiyoshi Sannai, Yusuke Iwasawa, Yutaka Matsuo
We present the group equivariant conditional neural process (EquivCNP), a meta-learning method with permutation invariance in a data set as in conventional conditional neural processes (CNPs), and it also has transformation equivariance in data space.
no code implementations • 27 Dec 2020 • Wataru Kumagai, Akiyoshi Sannai
However, universal approximation theorems for CNNs have been separately derived with individual techniques according to each group and setting.
no code implementations • 23 Oct 2020 • Yuuki Takai, Akiyoshi Sannai, Matthieu Cordonnier
The classical approach to measure the expressive power of deep neural networks with piecewise linear activations is based on counting their maximum number of linear regions.
no code implementations • 15 Oct 2019 • Akiyoshi Sannai, Masaaki Imaizumi, Makoto Kawano
To describe the effect of invariance and equivariance on generalization, we develop a notion of a \textit{quotient feature space}, which measures the effect of group actions for the properties.
no code implementations • 25 Sep 2019 • Akiyoshi Sannai, Masaaki Imaizumi
Learning problems with data that are invariant to permutations are frequently observed in various applications, for example, point cloud data and graph neural networks.
no code implementations • 17 Jun 2019 • Akinori Tanaka, Akiyoshi Sannai, Ken Kobayashi, Naoki Hamada
In this paper, we analyze the asymptotic risks of those B\'ezier simplex fitting methods and derive the optimal subsample ratio for the inductive skeleton fitting.
no code implementations • 5 Mar 2019 • Akiyoshi Sannai, Yuuki Takai, Matthieu Cordonnier
In this paper, we develop a theory about the relationship between $G$-invariant/equivariant functions and deep neural networks for finite group $G$.
no code implementations • 18 May 2018 • Akiyoshi Sannai
Furthermore, as as application of this theory, we prove that the loss functions can reconstruct the inputs of the training samples up to scalar multiplication (as vectors) and can provide the number of layers and nodes of the deep neural network.