Yuxi XIE
@sigrid_xie
Visiting Scholar @ucsbNLP | CS Ph.D. Student @wing_nus @NUSComputing | Ex Intern @MSFTResearch | Undergrad @PKU1898
Check out our poster session tomorrow at iclr.cc/virtual/2025/p…
🧑💻 Human software engineers constantly re-evaluate their approaches through experience. 🤖 However, LLM-based software agents can often get stuck in ineffective dead ends. Introducing SWE-Search: a multi-agent framework integrating search and self-refinement to enable software…
📣Hey #LLM processing folks (#NLProc , @iclr_conf folks), please RT and spread the word! Reasoning and planning are🔥topics in LLM. Consider your #ICLR2025 workshop schedule and join👇our workshop below & hear from our 🔑🗒️🔊s! See you over here in 🇸🇬!
🚀 Exciting News! Workshop on Reasoning and Planning for Large Language Models @ ICLR 2025 is coming 🌟 Please visit our official website: 👉 …shop-llm-reasoning-planning.github.io With the release of o1 Pro and the growing interest in research on reasoning and planning capabilities of LLMs,…
happy to see an approach that took inspiration from SWE-Search achieving SOTA on SWE-bench. while there is still plenty of room for improvement, search is a vital tool for navigating complex SWE environments, and I expect more approaches to follow suit.
👑
Interesting progress from @rm_rafailov and @DivGarg9 et. al following our work (applied to mathematical and commonsense reasoning): Monte Carlo Tree Search Boosts Reasoning via Iterative Preference Learning arxiv.org/abs/2405.00451 (Also discussed in Llama-3 paper, @AIatMeta )
Super excited to announce what we have been working on in the last six months - Agent Q is out now! This is a framework for self-supervised agent reasoning and search that can self-correct and autonomously improve by self-play and RL on real tasks on the real internet! 👇
Flying to #NeurIPS2024 tmr! Excited to connect with friends old & new. I'll be presenting the following works: 🪧[Poster] COrAL: arxiv.org/abs/2410.09675 🎙️[Lightning Talk] MCTS-DPO: arxiv.org/abs/2405.00451 🪧[Poster] DeGCG: arxiv.org/abs/2408.14866 Drop by and have a chat!
![sigrid_xie's tweet image. Flying to #NeurIPS2024 tmr! Excited to connect with friends old & new.
I'll be presenting the following works:
🪧[Poster] COrAL: arxiv.org/abs/2410.09675
🎙️[Lightning Talk] MCTS-DPO: arxiv.org/abs/2405.00451
🪧[Poster] DeGCG: arxiv.org/abs/2408.14866
Drop by and have a chat!](https://pbs.twimg.com/media/GeX4kD0bUAE14f9.jpg)
🚀 Excited to share our work @UCSB – 🪸 COrAL: Order-Agnostic Language Modeling for Efficient Iterative Refinement! 📌 Built on AR-LLM, COrAL unifies order-agnostic modelling and denoising, enabling efficient iterative refinement at inference time. 🔗 buff.ly/404vWsM
🚨😱Obligatory job market announcement post‼️🤯 I'm searching for faculty positions/postdocs in multimodal/multilingual NLP and generative AI! I'll be at #NeurIPS2024 presenting our work on meta-evaluation for text-to-image faithfulness! Let's chat! Website in bio, papers in🧵
🎉 Check out our works V-DPO and MVP-Bench, which will soon appear at #EMNLP Findings, presented by Yuxi @sigrid_xie 🗓️ buff.ly/3YZ2GTb 🚀 V-DPO enhances visual guidance in DPO to mitigate LVLM hallucination. See you in Miami!🏖️ 🔗 buff.ly/3CjUOTn 🧵 1/3
MVP-Bench🥇 buff.ly/3CnIMsp 🧠 Can LVLMs perceive both an image's overall semantics and its finer details? 🚀 We constructed MVP-Bench to investigate whether LVLMs perform differently across different levels of granularity. (1/4) 🔗 buff.ly/4hGip0W
Arrived at #EMNLP2024 #emnlp Excited to present our works of V-DPO, MVP-Bench, and DeGCG! 🤓 MVP-Bench: Session 06, Nov 13 (Wed) 10:30-12:00 DeGCG: Session 09, Nov 13 (Wed) 16:00-17:30 V-DPO: Session 12, Nov 14 (Thu) 14:00-15:30 See you in Miami! 🏖️
Our paper on “Self-Evaluation Guided Beam Search” has been accepted to #NeurIPS 2023 📷. Stay tuned for the upcoming Llama-2 backboned results on reasoning benchmarks. Please visit our project page at guideddecoding.github.io for the latest updates.
Check out our recent work (preprint: arxiv.org/abs/2305.00633) about LLM Reasoning by our Yuxi XIE @sigrid_xie (with guidance and collaboration of Min-Yen KAN @knmnyn, Junxian HE @junxian_he, Qizhe XIE @QizheXie, Kenji Kawaguchi, Yiran ZHAO @yiran_zhao924, and Xu ZHAO @xu_Zhao0).
#NLProc : the LREC-COLING '24 calls are out . All due 13 Oct 2023 AOE. Please R/T! 👉 🆕website at lrec-coling-2024.org for the paper, tutorial & workshop proposal calls (+author kit)! @LrecColing2024 @NicolettaCZ @AlexLenci1966 @axmum @roman_klinger @chokkanorg
📢The wait is over❣️LREC-COLING '24 (Torino 🇮🇹) 2nd call & 1st call for tutorial & workshop proposals out! All due📅 13 Oct 2023 AOE (no ARR needed) 👉 lrec-coling-2024.org (new URL) Pls submit your work & help spread news (RT 📨please)! @LrecColing2024 @NicolettaCZ