Toshinori Kitamura

Postdoc / The University of Alberta / Reinforcement Learning Theory

[Toshinori Kitamura @t_kitamura14]

I'm happy to share that our paper was accepted as a #NeurIPS25 spotlight! This is the first linear CMDP algorithm with sublinear regret and episode-wise safety.

Grateful to my amazing collaborators @omron_sinicx (esp. @Tdash_Koz) and @Arnobg32

https://t.co/LeUNyoGC4F

[Tadashi Kozuno @Tdash_Koz]

求職中です！日本から働ける研究ポジションをご存じでしたら、ぜひ教えてください！専門は強化学習、マルチエージェント強化学習、Learning in Gamesです。もともとは理論寄りの研究をしていましたが、現在はエージェントシステムの再帰的自己改善に興味があります。

[Taiji Suzuki @btreetaiji]

@matsui_kota そうですね，深層NNもTransformerも結局はカーネル法なのです．すべてはカーネル法のためにあるのです．

[Csaba Szepesvari @CsabaSzepesvari]

I can nominate one Canada Impact+ postdoc candidate at UAlberta. Please reach out only if you have a strong theory track record (learning theory/bandits-RL theory/LLM reasoning). Deadline: May 10 (23:59 MT. Email: CV + best papers + brief pitch. Pls RT https://t.co/pnRmnxPyCi

[Arnob Ghosh @Arnobg32]

This means that the sim-to-real gap is not just a matter of better tuning or larger-scale experiments. It reflects a deeper computational barrier in robust reinforcement learning. This work is led by @t_kitamura14
Please see here.. (n/n)
https://t.co/qAH6ROYnPI

[Alon Turing @chaumian]

Statistical Learning Theory in Lean 4: Empirical Processes from Scratch https://t.co/HI84oyJon6

[Nishioka @Nishi2357]

数理最適化に現れる様々な凸関数の一般化をまとめたサーベイ論文（プレプリント）
https://t.co/L4A5Ot1Q0n

[田中一敏 Kazutoshi TANAKA @sports_robots]

#クラウドファンディングCAMPFIRE にて、
研究者の業績管理を楽にするサービス PubListAuto の支援募集を開始しました。
国内外、大小問わず、論文等の自分の発表情報を網羅的にウェブから集約して、利用可能にするサービスです。
ご支援よろしくお願いします！
https://t.co/rz3LjsarOt

[Toshinori Kitamura @t_kitamura14]

@momiji_fullmoon 後で自分も気づきましたが同じ卓にいました👍

また挨拶します🙆‍♂️

[Toshinori Kitamura @t_kitamura14]

neurips会場付近でプロテインを買える場所はないのか

[Toshinori Kitamura @t_kitamura14]

@momiji_fullmoon お、ありがとうございます！
筋トレをします

[Toshinori Kitamura @t_kitamura14]

I’ll be presenting a poster on episode-wise safe exploration for linear CMDPs at #neurips25 .

See you in San Diego!

https://t.co/F1YE5CF7jd

[Max Simchowitz @max_simchowitz]

🧐🧐 Why do we pretrain LLMs with log likelihood? Why does action chunking work so well in robotics? Why is EMA so ubiquitous? And could their be a mathematical basis for Moravec’s paradox? 🤖🤖

Come check out our NeurIPS  2025 Tutorial “Foundations of Imitation Learning” with @canondetortugas and Adam Block, Tuesday 130-4pm, to find out! (🧵for details)

[Csaba Szepesvari @CsabaSzepesvari]

@karpathy @karpathy I think it would be good to distinguish RL as a problem from the algorithms that people use to address RL problems. This would allow us to discuss if the problem is with the algorithms, or if the problem is with posing a problem as an RL problem. 1/x

[Toshinori Kitamura @t_kitamura14]

@kato_jp おめでとうございます！サンディエゴで会いましょう🙆‍♂️

[Nishioka @Nishi2357]

「任意の停留点が大域的な最小点になる関数のクラス」であるinvex関数に関する論文を書きました。定義に現れるカーネル関数について色々調べています。具体例や図を多く載せたのでinvex関数の入門としてもいいと思います
https://t.co/1kCYQ0kDRV

[Nan Jiang @nanjiang_cs]

My 3rd blogpost on PG, the topic I am least familiar with but get asked a lot, so I thought I'd just put together the very limited stuff I know on this topic. Somehow the post gets cynical from time to time🙃

https://t.co/bD7coJJ7mI

[Arnob Ghosh @Arnobg32]

As AI is replacing human decision-making, it is important to sit back and consider the risk aspect of the cost associated with the decision. In this podcast, I talked about our recent ICML paper (https://t.co/h1SV8cHEsl) on risk-constrained MDP. Thanks @executive_code #AIrisk

[Toshinori Kitamura @t_kitamura14]

@nissymori1 あざます！頑張ります👍

[Toshinori Kitamura @t_kitamura14]

@IjongminS Thank you Jongmin! See you in conferences 👍

[Toshinori Kitamura @t_kitamura14]

Excited to share that I’ll be joining the University of Alberta as a postdoc next week! 🎉 I’ll be working on RL theory in Csaba Szepesvári’s lab at Amii!