Paper 🧵! w/ Samuel Stanton Anqi (Angie) Liu Suchi Saria #ICML2024 arxiv.org/abs/2405.06627 We study 2 Qs: 1) Can AI uncertainty quant. via #ConformalPrediction extend to any data distribution? 2) Are there practical CP algorithms for AI/ML agents 🤖w/ feedback-loop data shifts? 1/

🧐Looking for a super-fast 🚀 speech-to-speech translation (S2ST) system while maintaining high-quality 📈generation? Our paper uses ideas from diffusion models to improve S2ST. 💡✨ arXiv link: arxiv.org/abs/2405.13274 Github repo: github.com/steventan0110/… A thread 🧵

我们离理想的AutoGPT还有多远？[1/4] AutoGPT[1]已经163k star了，AutoGPT的开发者雕花了一年多，但它仍然停留在demo阶段，算不上产品（即使面向开发者)。这和传统开源软件的发展轨迹相差甚远，核心原因是Agent的上限由底座模型决定

Will be presenting my work on LLM-MT at #naacl today, at In-Person Poster Session 7 (11am). Please come if you are interested in machien translation and multilinguality! Would also love to discuss multimodal LLM (speech + text) and non-autoregressive decoding (NAT, diffusion)!

Here’s some better news: Combining CPO and SimPO can likely improve the model! Check out more details in our GitHub code: github.com/fe1ixxu/CPO_SI…

YESSSSSSS GEORGE TAKES POLE AT HIS HOME RACE!!!!❤️💙

Low-resource, high-risk: A Johns Hopkins University study (feat. Daniel Khashabi 🕊️, Lingfeng Shen, Weiting (Steven) Tan, Yunmo Chen, Jack Jingyu Zhang, Haoran Xu, & more) finds that #AI text generators are more likely to create harmful content in certain languages. Learn more: bit.ly/3Y6SacA