🚩(1/2) Please help forward the Call for the 2024 Adversarial Machine Learning (AdvML) Rising Star Awards! We promote junior researchers in AI safety, robustness, and security. Award events are hosted at AdvML'Frontiers workshop NeurIPS Conference 2024 Info: sites.google.com/view/advml/adv…

Joint work w/ Ane Zuniga, Xinwei Zhang, Johannes P. Dürholt, Payel Das, Jie Chen, Wojciech Matusik, and Mina Konaković Luković, a wonderful collaboration between MIT CSAIL, MIT-IBM Watson AI Lab, and Evonik. Code is also open-sourced github.com/yunshengtian/B…

Good morning, Vienna! ICML Conference Join us tomorrow for the LLMs and Cognition workshop at Lehar 3. We have an incredible lineup of speakers: Subbarao Kambhampati (కంభంపాటి సుబ్బారావు), Antoine Bosselut, Chelsea Finn, Ethan Gotlieb Wilcox, and Melanie Mitchell! Explore our schedule and accepted papers here: llm-cognition.github.io

Our LLMCog workshop is starting now with Subbarao Kambhampati (కంభంపాటి సుబ్బారావు)’s talk at Lehar 3 at Les McIsaac

After an awesome series of talks by Subbarao Kambhampati (కంభంపాటి సుబ్బారావు) Chelsea Finn Antoine Bosselut and Ethan Wilcox, and a very engaging panel, we are closing the workshop with the talk from amazing Melanie Mitchell. Thanks to all the speakers, panelists, poster presenters, audience, @ICML, IBM Research

Had some nice time with the workshop attendees before leaving #ICML2024 #Vienna

Really nice and insightful IEEE Signal Processing Society interview with the wise-beyond-her-years Celia Cintas. signalprocessingsociety.org/newsletter/202…

How many more of these tragic incidents must take place before we make the change …

Come to our #COLM2024 poster #13 this afternoon and hear from Irene Ko on test-time estimation robustness-accuracy trade-offs in LLMs with synthetic data. Joint work with Pin-Yu Chen Yung-Sung Chung Luca Daniel and myself.

Stop by at our poster today if you are interested in protein representation learning

We will be also at the MLSB workshop at #NeurIPS today to present the joint work with our amazing collaborators from #OxfordUniv, Lucy Vost and Charlotte Deane, on improving diffusion models towards generating more structurally plausible molecules mlsb.io/papers_2024/Im…

So good to see that our NOPE (no positional encoding) work from NeurIPS 2023 neurips.cc/virtual/2023/p… is now being utilized in Llama 4 models