「Schrödinger Bridge問題に基づく拡散生成モデル学習」という題目で講演します。基礎から最新の研究動向まで網羅的に解説します。他にも魅力的な講演がたくさんあるので、奮ってご参加ください！

先日の講演資料を公開しました。ML分野でのSchrödinger Bridge問題に関する研究がざっと把握できる内容になっています。 speakerdeck.com/takeshi_koshiz…

A Survey on Statistical Theory of Deep Learning: Approximation, Training Dynamics, and Generative Models. (arXiv:2401.07187v1 [stat.ML]) ift.tt/VHwOmbC

📢 In our new UAI 2025 paper, we do neural optimal transport with costs defined by a Lagrangian (e.g., for physical knowledge, constraints, and geodesics) Paper: arxiv.org/abs/2406.00288 JAX Code: github.com/facebookresear… (w/ A. Pooladian, C. Domingo-Enrich, Ricky T. Q. Chen)

Some related papers for our recent Lagrangian OT: 0. On amortizing convex conjugates for OT 1. Neural Lagrangian Schrödinger Bridge 2a. Deep Generalized Schrödinger Bridge 2b. DGSB Matching 3. Wasserstein Lagrangian Flows 4. Metric learning via OT A 🧵 summarizing these ❤️

階層的な構造を持つ最適輸送問題を、圏論的なアイディアで定式化したら、めちゃくちゃ爆速なアルゴリズムが作れた、という話をプレプリントとして公開しました: arxiv.org/abs/2408.08550

Our paper on mean-field analysis of Fourier neural operators has been accepted at NeurIPS 2024! Stay tuned for the updated version. arxiv.org/abs/2310.06379

Our paper on Neural Operator: 'Understanding the Expressivity and Trainability of Fourier Neural Operator: A Mean-Field Perspective' was accepted at #NeurIPS2024 🎉 🇨🇦 paper link: arxiv.org/abs/2310.06379 Huge thanks to Masahiro Fujisawa , Yusuke TANAKA , enjoy my life

以下の論文がICLR2025に採択されました！

以下の論文がICLR2025に採択されました！長谷川さん、M2で2年連続ICLRに採択されててすごい。

The following paper has been accepted to ICML 2025! Benign Overfitting in Token Selection of Attention Mechanism Keitaro Sakamoto, Issei Sato arxiv.org/abs/2409.17625

The following paper has been accepted to ICML 2025! On Expressive Power of Looped Transformers: Theoretical Analysis and Enhancement via Timestep Encoding Kevin Xu, Issei Sato arxiv.org/abs/2410.01405

🎉 Our paper on "the expressive power of Looped Transformers" was accepted at #ICML2025 ! To the best of our knowledge, this is the first study to analyze their function approximation capabilities, including approximation rates and universality. arxiv.org/abs/2410.01405

Currently visiting ZIB in Berlin

1/ Where do Probabilistic Models, Sampling, Deep Learning, and Natural Sciences meet? 🤔 The workshop we’re organizing at #NeurIPS2025! 📢 FPI@NeurIPS 2025: Frontiers in Probabilistic Inference – Learning meets Sampling Learn more and submit → fpiworkshop.org

We’re announcing a major advance in the study of fluid dynamics with AI 💧 in a joint paper with researchers from Brown University, New York University and Stanford University.

Solving control problems can be hard. This is why we introduce trust region methods, approaching them iteratively in a systematic way. In fact, this can be understood as a geometric annealing from prior to target with adaptive steps. More at NeurIPS Conference, arxiv.org/pdf/2508.12511.