Furong Huang
@furongh
Associate professor of @umdcs @umiacs @ml_umd at UMD. Researcher in #AI/#ML, AI #Alignment, #RLHF, #Trustworthy ML, #EthicalAI, AI #Democratization, AI for ALL.
ID: 195674678
https://furong-huang.com/ 27-09-2010 09:11:38
1,1K Tweet
5,5K Followers
2,2K Following
It was a lot of fun catching up with folks at MSR-NYC! Thanks for the invitation John Langford!
Can we train vision-based reinforcement learning agents that achieve super-human performance against human players in racing games? Check out our latest work to be presented at RL_Conference! Paper: rlj.cs.umass.edu/2024/papers/Pa… Website: ai.sony/publications/A… A very fast 🧵(1/7)
Token-wise MDPs/RL has recently found lots of interest in alignment, including - Our controlled decoding (arxiv.org/abs/2310.17022) -Souradip Chakraborty's transfer Q-star (arxiv.org/abs/2405.20495) -Xiangyu Qi's token-wise constrained fine-tuning (arxiv.org/abs/2406.05946)
Third keynote by Andy Barto RL_Conference , arguing that it was always RL, with a standing ovation at the end!
Andy Barto answering a question RL_Conference that I have been thinking about for 2 years. Turns out ML (supervised learning) is a special case of RL we constructed for some reason...
Really nice initiative by Ben Eysenbach , who prepared these posters (hung around RL_Conference ) of notable women in RL !
Excited to share the slides of the 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐦𝐨𝐝𝐞𝐥 𝐢𝐧𝐟𝐞𝐫𝐞𝐧𝐜𝐞: 𝐭𝐡𝐞𝐨𝐫𝐲 & 𝐚𝐥𝐠𝐨𝐫𝐢𝐭𝐡𝐦𝐬 tutorial that Ahmad Beirami and I gave at #ISIT2024. We covered basics of 𝘪𝘯𝘧𝘦𝘳𝘦𝘯𝘤𝘦, 𝘢𝘭𝘪𝘨𝘯𝘮𝘦𝘯𝘵 and 𝘦𝘧𝘧𝘪𝘤𝘪𝘦𝘯𝘤𝘺 from a theoretical lens.
