Sho Sonoda

Overview

• AI4Math

  • Theorem Proving in Lean + LLM（理研AIP数学系セミナー2025）slide

  • （一般向け）AIにおける自動証明について（数学セミナー増刊 2025.06）補足資料

• Deep Learning Theory

  • Deep Ridgelet Transform: Harmonic Analysis for Deep Neural Networks: slide v. Jul2025 (english) 

  • 積分表現でニューラルネットを理解する（OCAMI「人工知能と数学」）YouTube動画, slide v2024年3月（日本語）

  • （一般向け）機械学習を支える数学（数学キャラバン2024）slide

  • AIP Open Seminar recordings, Apr2021 (YouTube video, english)

Specific Topics

• AI4Math

  • LeanConjecturer: LLM Conjecturing for AI Theorem Proving: slide v. Jul2025 (english)

  • Lean Formalization of Generalization Error Bounds by Rademacher Complexity: v. Apr2025 (english)

    • Lean4 による汎化誤差評価の形式化: v. 2025年6月（日本語 by 塚本さん）

• Deep Learning Theory

  • Growth Rate Controls Depth Dependence of Generalization Error Bounds: slide v. Jun2025 (english)

• Ridgelet Analysis

  • Deep Ridgelet Transform for Joint-Group-Equivariant Machine: slide v. Jul2025 (english)

  • General Ridgelet Transform Induced from Group Invariant Functions: slide v. Dec2023 (english)

  • Ridgelet Transform for Group Convolutional Neural Networks: slide v. Nov2022 (english) v. 2022年9月（日本語）, poster

  • Ridgelet Transform for Neural Networks on Noncompact Symmetric Space and Helgason-Fourier Analysis: slide v. Aug2022 (english), poster

  • Ghosts (or Null Space) in Neural Networks: slide v. Jun2021 (english)

    • ニューラルネットの零空間（勉強会資料）：slide v2021年10月（日本語）

  • Approximation Lower Bounds for Neural Networks: slide v. Sep2020 (english) poster v. Jul2023 (english)

• Transport Analysis (Neural ODE)

  • Transport analysis for denoising autoencoders: slide v. Aug2019 (english)

    • 積分表現理論と輸送理論（ICML2019読み会資料） : slide v2019年8月 (日本語)

• Quantum Machine Learning

  • Quantum algorithm for optimal random Fourier features: slide v. Nov2021 (english)

Research Interest

• Theory of Neural Networks

  • Integral representation of neural networks, Ridgelet analysis

  • Transportation analysis of deep neural networks, Wasserstein geometry

  • Synthesizing method for neural networks, Probabilistic numerics

  • Generalization error analysis from the perspective of Geometric group theory

• Machine Learning, Data Science

  • Non-parametric statistics, kernel methods

  • Particle filtering, EEG signal processing

• AI4Math

  • Automated Theorem Proving (ATP)

• Quantum Machine Learning (QML)

• Autonomous Driving

Contact

sho.sonoda [at] riken.jp (most preferable)

sho.sonoda [at] aoni.waseda.jp (expired at the end of April 2018)

s.sonoda0110 [at] toki.waseda.jp (permanently valid alumni address)