Xuehang Guo

[Xuehang Guo @XuehangG]

Check out our poster at COLM 2027 (poster #119)!

[Xuehang Guo @XuehangG]

🧵 [5/5]

SFT and RL validate AnaBench and Anagent, achieving notable gains in scientific table & figure analysis.

See more details in:
🌐 Homepage: https://t.co/026sctE6vT
💻 Code: https://t.co/cRKmR4DtSR
🤗 Data: https://t.co/MqkOLgu6tM
📄 Paper: https://t.co/rIlcIelDjj

[Xuehang Guo @XuehangG]

Very glad to share that our paper has been accepted to EMNLP 2026! My greatest thanks to my exceptional mentors Qingyun @eagle_hz , Tom @Hoper_Tom , and Zhiyong.

🌐 Homepage: https://t.co/026sctE6vT
📄 Paper: https://t.co/rIlcIelDjj

[Xuehang Guo @XuehangG]

🧵 [4/5]

ANAGENT mirrors how researchers actually work, with four specialized agents:

🎯 PLANNER decomposes the task
🔍 EXPERT retrieves evidence
🖋️ SOLVER writes the analysis
🔧 CRITIC reviews and refines

[Xuehang Guo @XuehangG]

🧵 [2/5]

ANABENCH: a scientific analysis benchmark across 9 scientific domains and 170 fine-grained subdomains.

Tables, figures, and combined multimodal inputs, all built from real research papers.

[Xuehang Guo @XuehangG]

Prof. Wang @zhenhailongW at Westlake University is recruiting students! Highly recommend! 🙌

Prof. Wang is very expert in multimodal research, and is incredibly kind and supportive to all his students. If you’re applying for 27Fall, definitely consider applying! :)

[Zhenhailong Wang @zhenhailongW]

Excited to share that I’ll join Westlake University as a tenure-track AP in AI from summer 2027! Welcome to join the lab as a "co-founder"! I am looking for: 2–3 PhD students (Fall 2027), RAs/interns. Check out the recruiting page for more details https://t.co/rMYwM2FyRJ :)

[Yi R. (May) Fung @May_F1_]

Excited crossing 4K citations! 🥳

Come check out our ACL'26 oral presentation of 𝗖𝗶𝘁𝗲𝗚𝘂𝗮𝗿𝗱 7/4th, where Kath @yeemanchoi wil share more on 𝘴𝘢𝘧𝘦𝘨𝘶𝘢𝘳𝘥𝘪𝘯𝘨 𝘵𝘩𝘦 𝘷𝘢𝘭𝘪𝘥𝘪𝘵𝘺 𝘰𝘧 𝘴𝘤𝘪𝘦𝘯𝘵𝘪𝘧𝘪𝘤 𝘤𝘭𝘢𝘪𝘮 𝘳𝘦𝘧𝘦𝘳𝘦𝘯𝘤𝘦𝘴 & what's next to come~📢

[Xuehang Guo @XuehangG]

Looking for citations you can actually trust? Check out our work CiteGuard for faithful citation attribution via retrieval-augmented validation!

Kath @yeemanchoi will present on ACL, Sunday, 7/5, 16:00–17:30, in Harbor B–C!

Paper: https://t.co/GVNEFdby4P

[Qingyun Wang @eagle_hz]

Come check out our ACL oral on Sunday, 7/5, 16:00–17:30, in Harbor B–C!
Kath @yeemanchoi will present our work on CiteGuard 🛡️: a citation bodyguard for LLMs 📚
Paper: https://t.co/2gCTbJ0jTq

[Yi R. (May) Fung @May_F1_]

We need more 𝗼𝗽𝗲𝗻, 𝗿𝗲𝗮𝗹𝗶𝘀𝘁𝗶𝗰 𝗮𝗴𝗲𝗻𝘁 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝘀 for training and evaluating agents! 💡

🚀But what are the 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝗲𝗻𝘃𝗶𝗿𝗼𝗻𝗺𝗲𝗻𝘁𝘀 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱?

📈What are the 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗯𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸𝘀 for these environments in training and evaluation, and how can we 𝘀𝗰𝗮𝗹𝗲 𝘂𝗽 the number of available environments?

🌟Most importantly, how should we utilize these environments: 𝗥𝗟 𝗼𝗿 𝗯𝗲𝘆𝗼𝗻𝗱?

If you’re interested in discussing together, come join us at our workshop on “𝙎𝙘𝙖𝙡𝙞𝙣𝙜 𝙀𝙣𝙫𝙞𝙧𝙤𝙣𝙢𝙚𝙣𝙩𝙨 𝙛𝙤𝙧 𝘼𝙜𝙚𝙣𝙩𝙨” @NeurIPSConf tmr (7th Dec)! We have an amazing lineup of invited speakers and panelists, including 𝐄𝐝𝐰𝐚𝐫𝐝 𝐆𝐫𝐞𝐟𝐞𝐧𝐬𝐭𝐞𝐭𝐭𝐞 from 𝐆𝐨𝐨𝐠𝐥𝐞 𝐃𝐞𝐞𝐩𝐌𝐢𝐧𝐝 and 𝐒𝐡𝐮𝐲𝐚𝐧 𝐙𝐡𝐨𝐮 from 𝐃𝐮𝐤𝐞.

Also check out our latest 𝐬𝐮𝐫𝐯𝐞𝐲 𝐩𝐚𝐩𝐞𝐫 on the topic led by Yuchen Huang: https://t.co/gvsdqkp6jf 🎯

[DailyPapers @HuggingPapers]

Dive deeper into PAPO on Hugging Face!

This framework improves multimodal reasoning without extra data or external reward models.

Explore the paper, models, and datasets:
Paper: https://t.co/dV9WqD3SDi
Models: https://t.co/0haxntZnzl
Data: https://t.co/S7cL5lvxWO

[Haiyang Xu @xuhaiya2483846]

🔥🔥🔥We introduce Perception-Aware Policy Optimization for multimodal reasoning.
✅ +4.4% overall on multimodal benchmarks
✅ +8.0% on highly vision-dependent tasks
✅ −30.5% perception errors
🌐 Project: https://t.co/iiTArVTwQA
📄 Paper: https://t.co/Mh4F4kbp07
@_akhaliq

[Zhenhailong Wang @zhenhailongW]

Learning to perceive while learning to reason!
We introduce PAPO: Perception-Aware Policy Optimization, a direct upgrade to GRPO for multimodal reasoning. PAPO relies on internal supervision signals. No extra annotations, reward models, or teacher models needed. 🧵1/3

[Xuehang Guo @XuehangG]

Great thanks to my exceptional collaborators and mentors! @xingyaow_, @YangyiChen6666, @ZoeyLi20, @Glaciohound, @ManlingLi_, @hengjinlp

[Xuehang Guo @XuehangG]

Many thanks!
Please check out our paper and project website for more details:
Paper: https://t.co/jsJVkpWLEC
Website: https://t.co/pPZHGQc1wA

[Xuehang Guo @XuehangG]

As software engineering becomes increasingly collaborative, how do all coding agents keep in sync with evolving environments?
Introducing our latest work:
SyncMind: Measuring Agent Out-of-Sync Recovery in Collaborative Software Engineering.
🌐Website: https://t.co/pPZHGQc1wA

[Xuehang Guo @XuehangG]

I extend my huge gratitude to my co-authors: Xingyao Wang @xingyaow_, Yangyi Chen @YangyiChen6666, Zoey Li @ZoeyLi20, Han Chi @Glaciohound, Manling Li @ManlingLi_, Heng Ji @hengjinlp