📢 New @neuripsconf.bsky.social paper with @jan-peters.bsky.social & @olegarenz.bsky.social : "A Geometric Perspective on Reward Function Updates in Inverse Reinforcement Learning" We show that natural-gradient reward updates in inverse RL require no Fisher matrix at all 🧵👇

Australopithecus 3 million years ago:

josh (oldfriend99)@oldfriend99.bsky.social · 2w ago

You guys should check out Tool

I had a lovely time in Seoul presenting Trust Region Inverse RL (arxiv.org/pdf/2605.11020) at @icmlconf.bsky.social! Also got a chance to visit and give a talk at the Robot Learning Lab at SNU (rllab.snu.ac.kr). They’re doing amazing work on RL for robotics 🦾

📢 Excited to share that our paper 𝗧𝗿𝘂𝘀𝘁 𝗥𝗲𝗴𝗶𝗼𝗻 𝗜𝗻𝘃𝗲𝗿𝘀𝗲 𝗥𝗲𝗶𝗻𝗳𝗼𝗿𝗰𝗲𝗺𝗲𝗻𝘁 𝗟𝗲𝗮𝗿𝗻𝗶𝗻𝗴 (𝗧𝗥𝗜𝗥𝗟) has been accepted at @icmlconf.bsky.social 2026! 🌐 Project page: lnkd.in/dqNi68Mk 📄 Paper: lnkd.in/dzbcCbjU #ICML2026 #ReinforcementLearning #InverseRL #ImitationLearning #Robotics

Remember folks: if you have any small amount of artistic talent, any sex appeal, or can make your friends laugh, the richest man on earth is crying screaming throwing up because he can’t be you.

When plastic explosives go off nearby, do your ears ring at middle c?

It’s hilarious how I’m continually an idiot in hindsight

I would like to coin a new definition. 𝐇𝐨𝐛𝐛𝐲 𝐏𝐚𝐫𝐚𝐥𝐲𝐬𝐢𝐬 :⁣ 𝘞𝘩𝘦𝘯 𝘺𝘰𝘶 𝘩𝘢𝘷𝘦 𝘢 𝘴𝘶𝘥𝘥𝘦𝘯 𝘪𝘯𝘧𝘭𝘰𝘸 𝘰𝘧 𝘸𝘢𝘺 𝘵𝘰𝘰 𝘮𝘢𝘯𝘺 𝘤𝘳𝘦𝘢𝘵𝘪𝘷𝘦 𝘪𝘥𝘦𝘢𝘴 𝘣𝘶𝘵 𝘯𝘰𝘵 𝘦𝘯𝘰𝘶𝘨𝘩 𝘵𝘪𝘮𝘦 𝘰𝘳 𝘴𝘬𝘪𝘭𝘭 𝘧𝘰𝘳 𝘢𝘯𝘺 𝘰𝘧 𝘵𝘩𝘦𝘮. 𝘉𝘰𝘯𝘶𝘴 𝘱𝘰𝘪𝘯𝘵𝘴 𝘪𝘧 𝘺𝘰𝘶 𝘤𝘢𝘯𝘯𝘰𝘵 𝘥𝘦𝘤𝘪𝘥𝘦 𝘸𝘩𝘪𝘤𝘩 𝘰𝘯𝘦 𝘵𝘰 𝘣𝘦𝘨𝘪𝘯 𝘱𝘶𝘳𝘴𝘶𝘪𝘯𝘨.

🎞️ Ilford FP4+

📢 New blog post! 📢 I think partial observability in decision making is a fascinating topic. It leads to a ton of roadblocks that “traditional” planning/learning agents are just not equipped to handle. #MLSky anishhdiwan.github.io/notes/pomcp....

I hope there comes a day when ratm’s and soad’s music stops being globally relevant. I’m afraid that’s a pipe dream.

“I wonder what’s happening today in the allegedly most developed country in the world,” I say, logging on and immediately regretting it.

Random alley in Venice (from last summer) 🎞️ Portra 160 #filmphotography

NYC on a rainy evening 🎞️ (shot on Ultramax 400)

I have a new blog on machine learning and robotics! It’s an academia focused archive of my lecture notes, paper summaries, and misc. write ups. I’ve recently written posts on imitation learning and the policy gradient theorem (RL). #MLSky #Robotics anishhdiwan.github.io/notes/

Hey, the internet! I’ll try to keep the memes to a minimum :3 No promises though