Mohit Bansal

Parker Distinguished Prof @UNC. PECASE/ACL/AAAI Fellow. Director (@unc_ai_group). Past @Berkeley_AI @TTIC_Connect @IITKanpur #NLP #CV

[Mohit Bansal @mohitban47]

Deeply happy and honored to be elected as an ACL Fellow -- and to be a part of the respected cohort of this+past years' fellows (congrats everyone)! 🙏

All the credit (and sincere gratitude) to all my amazing students, postdocs, collaborators, mentors, and family! 🤗💙

[Mohit Bansal @mohitban47]

...

[Mohit Bansal @mohitban47]

👋 Looking forward to attending #ACL2026 (in-person in San Diego) and #ICML2026 (probably virtually) for these presentations/workshop keynotes & meeting everyone (also, I'll be in the Bay Area beforehand, for a keynote at the Apple Reasoning and Planning Workshop)!

Feel free to ping if you want to meet up in Bay/SD (I also have July1 partly free in SJ/SF, and several days in SD), and discuss research, life, etc. (we're also hiring at all levels: phd, postdoc, faculty)! 🙂

PS. also meet several of our awesome students/postdocs/alumni attending these 2 conferences to present these works.
👇👇

[Duy Nguyen @duynguyen772]

I will be presenting "GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs" tomorrow (July 7th) at #ACL2026 Poster Session G (Grand Hall, 11:00 - 12:30). Drop by to chat about the paper and other topics in safety/interpretability!

[Jaehong Yoon @jaeh0ng_yoon]

Excited to be in Korea for #ICML2026! I’ll be giving a few talks, organizing the SCALE workshop, and attending several events throughout the week.

Looking forward to meeting everyone. Feel free to reach out if you’ll be around the conference venue 😁

[Mohit Bansal @mohitban47]

Attending authors of these works at @icmlconf -->

Justin Chih-Yao Chen @cyjustinchen
Hyunji Amy Lee @hyunji_amy_lee
Hanqi Xiao @hanqi_xiao
Jaemin Cho @jmin__cho (now faculty at @jhuclsp @JHUCompSci )
Jaehong Yoon @jaeh0ng_yoon (now faculty at @NTUsg),
and our several collaborators!

[Mohit Bansal @mohitban47]

If you are at #ICML2026, check out these presentations and meet the presenters! 🙂

-- skill-inference in mixture-of-experts for heterogeneous reasoning

-- generalized correctness prediction models

-- effective reasoning and intrinsic dimensionality (spotlight, top-2.2%)

-- efficient video camera control learning

-- multimodal fact-level attribution for verifiable reasoning

-- scientific video reasoning in large multimodal models

-- Keynote: “Efficient Control, Memory, & Skill Planning in Multimodal Generative Agents” (virtual)
ICML Workshop on Scalable Learning & Optimization for Efficient Multimodal AI (SCALE)

👇👇

[Mohit Bansal @mohitban47]

👋 Looking forward to attending #ACL2026 (in-person in San Diego) and #ICML2026 (probably virtually) for these presentations/workshop keynotes & meeting everyone (also, I'll be in the Bay Area beforehand, for a keynote at the Apple Reasoning and Planning Workshop)!

Feel free to ping if you want to meet up in Bay/SD (I also have July1 partly free in SJ/SF, and several days in SD), and discuss research, life, etc. (we're also hiring at all levels: phd, postdoc, faculty)! 🙂

PS. also meet several of our awesome students/postdocs/alumni attending these 2 conferences to present these works.
👇👇

[ACL 2026 @aclmeeting]

There will be 3 honorable mentions and 1 winner presented by @adveisner & Mohit Bansal: https://t.co/069e4avZQT! #ACL2026NLP

[ACL 2026 @aclmeeting]

Don't forget to attend the "ACL Computational Linguistics Doctoral Dissertation Award 2026" session on Mon Jul 6, 9:009:30 AM (Location: Promenade) in the "Session 8: Oral/Poster Session D" (1/2) #ACL2026NLP

[Daeun Lee @danadaeun]

I will be presenting "VideoRepair: Self-Correcting Text-to-Video Generation with Misalignment Detection and Localized Refinement" at #ACL2026 today July 5 (2:00 - 3:30) at Grand Hall Poster Section B.

Looking forward to discussing and sharing thoughts about video reasoning/generation/understanding! 📽️

[Jaewoo Lee @jwlee8877]

I will be presenting "PRInTS: Reward Modeling for Long-Horizon Information Seeking" at #ACL2026 today July 5 (4:00 - 5:30). Looking forward to discussing and sharing thoughts in this area!

[Sapir Harary @ ACL 2026 @SapirHarary]

🚨 Presenting PrefixNLI (Oral) tomorrow at #ACL2026! 🌴
Hallucinations start with a single token. So why wait for a full sentence to detect them?

We extend Natural Language Inference to arbitrary text prefixes, detecting factual inconsistencies as soon as they arise and using that signal to guide decoding.

Our controlled decoding method delivers major faithfulness gains in summarization. A 3B model matches an 8B model in faithfulness while maintaining similar runtime, using half the memory, and is substantially faster than prior lookahead methods.

@SapirHarary, @hirscheran, @lovodkin93, @meetdavidwan, @mohitban47, Ido Dagan

📍 Regatta
⏰ 2:10pm

https://t.co/YHmGvbXkMD
Hope to see you there! 👋

#NLProc #ACL2026

[Tianyi Niu @niu_tianyi]

Just landed in San Diego for #ACL2026! I will be presenting our most recent work "Routing with Generated Data" at an oral session Tuesday July 7 (11:00 - 12:30, Harbor D). Looking forward to meeting new people and discussing interesting ideas!

[Qianqi "Jackie" Yan @qianqi_yan]

That's a wrap on the 4th Workshop on Advances in Language and Vision Research (ALVR) at #ACL2026 in San Diego! 🌊
47 submissions → 29 accepted papers, 6 spotlights, 5 invited talks, and a packed room all day. 🧵

[Qianqi "Jackie" Yan @qianqi_yan]

The workshop is going on right now, come join us at Harbor H&I!

[Mohit Bansal @mohitban47]

Landed in San Diego for #ACL2026 yesterday, with a pretty nice room view! 🏖️🌊

I will be at these 2 locations for keynotes today if you want to chat 👇

-- “Unifying Video and Audio for Multimodal Understanding and Generation”
23rd International Conference on Spoken Language Translation (@iwslt)
9.40-10.30am (https://t.co/aAgvMbpMQn)

-- “Long-Horizon Video Reasoning and Generation”
4th ACL Workshop on Advances in Language and Vision Research (#ALVR2026)
2.10-2.50pm (https://t.co/ypyriOvK2Q)

[Mohit Bansal @mohitban47]

👋 Looking forward to attending #ACL2026 (in-person in San Diego) and #ICML2026 (probably virtually) for these presentations/workshop keynotes & meeting everyone (also, I'll be in the Bay Area beforehand, for a keynote at the Apple Reasoning and Planning Workshop)!

Feel free to ping if you want to meet up in Bay/SD (I also have July1 partly free in SJ/SF, and several days in SD), and discuss research, life, etc. (we're also hiring at all levels: phd, postdoc, faculty)! 🙂

PS. also meet several of our awesome students/postdocs/alumni attending these 2 conferences to present these works.
👇👇

[Antonis Anastasopoulos @anas_ant]

Join us tomorrow at @iwslt for an exciting day (1 of 2) around spoken language translation:

https://t.co/bzsCohnhht

- keynote by @mohitban47 multimodal understanding and generation
- The findings of 7 Shared Tasks
- 7 oral presentations
- a poster session
#ACL2026nlp

[Elias Stengel-Eskin @EliasEskin]

Looking forward to #ACL2026, I'll be around July 3-8. Feel free to ping if you want to catch up!

I won't make it to #ICML2026, but many of my co-authors will be there and at ACL to share our work on:

-- calibrated correctness models/introspection
-- skill-based model routing
-- annotation-free routing
-- long-horizon info-seeking PRMs
-- attributed steering for VLMs, multimodal citations/attribution

🧵

[Justin Chih-Yao Chen @cyjustinchen]

Will be at #ICML2026 next week to present Skill-MoE:
📅 Jul 9 2:30 PM – 4:15 PM KST
📌 Hall A #1410

Excited to reconnect with old friends and meet new folks! My DMs are open if you’d like to grab a coffee ☕

🌐https://t.co/qYwVlLqoJy

[Mohit Bansal @mohitban47]

@aclmeeting @danadaeun @jwlee8877 @duynguyen772 @niu_tianyi @EliasEskin @UTCompSci @icmlconf @cyjustinchen @hyunji_amy_lee @hanqi_xiao @jmin__cho @jhuclsp @jaeh0ng_yoon @NTUsg @ArchikiPrasad @codezakh @gentaiscool @zhan1624 @SapirHarary @hirscheran @lovodkin93 @meetdavidwan @biunlp @HanWang98 @Xiaodong_Yu_126 @Wu_Jialian @XimengSun1 @AIatAMD @ArieCattan @paul_roit @byryuer @roeeaharoni @ISzpektor @sukwon_yun @TianlongChen4 @vaidehi_patil_ @ZunWang919 @JialuLi96 @hanlin_hl @ZiyangW00 @mandarjoshi_ @kentonctlee @ptshaw2 #ICML2026

-- SciVideoBench: Benchmarking Scientific Video Reasoning in Large Multimodal Models

@dengandong1227 @TaojiannanYang @shoubin621 Lincoln Spencer @mohitban47 Chen Chen @yeung_levy @XiaohanWang96

https://t.co/01hPVMnjKk

https://t.co/1VGX28VRRz

[Shoubin Yu @shoubin621]

🚨 New Paper Alert! Introducing SciVideoBench — a comprehensive benchmark for scientific video reasoning!

🔬SciVideoBench:

1. Spans Physics, Chemistry, Biology & Medicine with authentic experimental videos.

2. Features 1,000 challenging MCQs across three reasoning types: Conceptual, Hypothetical, and Quantitative.

3. Explicitly targets higher-order multimodal cognitive skills in LMMs.

4. Provides valuable insights on Chain-of-Thought prompting, scaling effects, and reasoning limits.
🧵

[Mohit Bansal @mohitban47]

👋 Looking forward to attending #ACL2026 (in-person in San Diego) and #ICML2026 (probably virtually) for these presentations/workshop keynotes & meeting everyone (also, I'll be in the Bay Area beforehand, for a keynote at the Apple Reasoning and Planning Workshop)!

Feel free to ping if you want to meet up in Bay/SD (I also have July1 partly free in SJ/SF, and several days in SD), and discuss research, life, etc. (we're also hiring at all levels: phd, postdoc, faculty)! 🙂

PS. also meet several of our awesome students/postdocs/alumni attending these 2 conferences to present these works.
👇👇

[Mohit Bansal @mohitban47]

#ICML2026

-- Effective Reasoning Chains Reduce Intrinsic Dimensionality (Spotlight top-2.2%)

@ArchikiPrasad , @mandarjoshi_ , @kentonctlee , @mohitban47 , @ptshaw2 @GoogleDeepMind

https://t.co/04AqSJF9Sr

https://t.co/IHz4s8U1La

[Archiki Prasad @ArchikiPrasad]

🚨Excited to share our new work viewing reasoning strategies as teaching tools: for fixed target model, which CoT strategies best support learning and generalization?

✨Our answer is intrinsic dimensionality (minimum effective capacity a model needs to solve the task).

Somewhat counterintuitively, adding CoT – which requires generating longer and more structured outputs – can reduce learning complexity. Good reasoning compresses the task, i.e., it reduces the degrees of freedom the model needs to map inputs to correct solutions.

🧵⬇️ (1/5)

[hyunji amy lee ✈️ ICML @hyunji_amy_lee]

I’ll be attending #ICML2026 in person and is looking forward to catching up with old friends and meeting new people! I’ll be presenting:

📌 Generalized Correctness Models: Learning Calibrated and Model-Agnostic Correctness Predictors from Historical Patterns
🗓️ Jul 7, 10:30 AM–12:15 PM • Hall A #2003

📌 Multimodal Fact-Level Attribution for Verifiable Reasoning
🗓️ Jul 8, 10:30 AM–12:15 PM • Hall A #1707

📌 Can Large Language Models Keep Up? Benchmarking Online Adaptation to Continual Knowledge Streams
🗓️ CATS Workshop, Jul 10 • Room 327

I’d love to chat about memory systems, multi-agent systems, and continual learning! Feel free to reach out if you'd like to chat 😄