Hi everyone! I’m attending #NeurIPS2025 in San Diego! Looking forward to connecting with folks across the AI community, and sharing our exciting work on AI for Science at #CFN#BrookhavenNationalLab
Excited to be in Montreal this week for COLM 2025!
I have two papers accepted this year:
1️⃣ The Unlearning Mirage, a dynamic framework for evaluating LLM unlearning.
2️⃣ With @Harsh_N_Lalai10, using 20 Questions as a creative framework to systematically evaluate geographical bias in LLMs.
Check out the individual threads!
"A 6th mass biological extinction is unfolding. We desperately need to safeguard rare, range-restricted species. Conserving a little more than 1% of land could do that." - Rodolfo Dirzo on a @FrontScience study mapping key biodiversity hotspots to protect: https://t.co/qhMD3JgNDl
Key insight 2, in presence of noisy supervision, a large pool of negatives is needed in loss. This is a semi-known thing but a large pool of negatives comes with huge GPU memory bottleneck in DE training. We solve this using a neat trick gradient caching (https://t.co/1m1esGg8za)
My completed drawing of Eric Cantona! ✍🏼
Right in time for his 56th birthday, Happy Birthday King! 👑
Would really appreciate all the retweets and comments friends! 🙏🏼😊
@ManUtd
Finally, the @cristiano artwork is now complete! 120+ hours of hardwork over many months, safe to say, it was all worth it in the end. ✍🏼🙌🏼
Would mean the world to me if you can show some love by liking and retweeting this! 🔄😊
@ManUtd#MUFC#UnitedFanArt#CR7
Gao, Xiang, and Shray Mathur. "Predictability of U.S. Regional Extreme Precipitation Occurrence Based on Large-Scale Meteorological Patterns (LSMPs)", #JClimate 34, 17 (2021): 7181-7198, https://t.co/rgik1080Vp