Yaoyao(Freax) Qian

🔬Building | 🐱🐶🐯🦁🦋

[Yaoyao(Freax) Qian @RubyFreax]

🥰Now is a #CoRL2024 Paper!

[Yaoyao(Freax) Qian @RubyFreax]

❓ How do you solve grasping problems when your target object is completely out of sight?

🚀 Excited to share our latest research! Check out ThinkGrasp: A Vision-Language System for Strategic Part Grasping in Clutter.

🔗 Site: https://t.co/bWNXlTiUea

[Hokin Deng @DengHokin]

#ICML Two last-authored papers at ICML 2026 this week.

A Very Big Video Reasoning Suite.

Video models today are optimized for one thing: visual quality. Whether they can reason, track causality, follow constraints, execute multi-step plans, has been nearly impossible to study, because the field never had the data. VBVR changes that.

Over 50 researchers and engineers across 30+ institutions built:

1) 150 reasoning tasks grounded in cognitive and developmental psychology
2) 2,015,000 samples and over one million videos, roughly 1,000× larger than the nine prior video-reasoning datasets combined
3) every task procedurally generated with deterministic ground truth, so evaluation is fully rule-based and verifiable, no LLM judges, and it agrees with human judgments

Insights:
1) Fine-tuning an open-source 14B model on VBVR lifts overall performance by 84.6 percent, a new state of the art that surpasses every proprietary system we evaluated, including Sora 2 and Veo 3.1.
2) The learned skills generalize to out-of-domain task families the model never saw, early evidence of emergent generalization in video reasoning.
3) And the honest part: gains saturate with scale, on two different architectures. A persistent gap to human performance remains that data alone cannot close. The bottleneck is architectural.
4) One principle kept surfacing throughout: controllability, doing exactly what is asked — is the bedrock of verifiable video reasoning.

Great thanks to my collaborators @MaijunxianWang @Rui147000038622 @juyi_lin @SaraJiran @thwiedemer @Kaiagaoqy @carrot0817_ @RubyFreax Lianyu Huang Zelong Hong Jiahui Ge Qianli Ma Hang He Yifan Zhou Lingzi Guo Lantao Mei Jiachen li Hanwen Xing Tianqi Zhao Fengyuan Yu Weihang Xiao Yizheng Jiao Jaden Hou Danyang Zhang Pengcheng Xu Boyang Zhong Zehong Zhao Gaoyun Fang John Kitaoka Yile Xu Hua Xu Kenton Blacutt @tin_ng_qn @siyuansong_ @HaoranSunMr Shaoyue Wen @LinyangNeuroAI @RunmingW @Yanzhi_Wang_ @Mengyue_Yang_ @ziqiao_ma @raphaelmilliere @fredahshi Nuno Vasconcelos @DanielKhashabi Alan Yuille @du_yilun @ZimingLiu11 @Brian_Bo_Li Dahua Lin @liuziwei7 @Vikashplus Yijiang Li @drowsyleilei @caizhongang

The second paper — "Vision Language Models Cannot Reason About Physical Transformation" — asks a question Piaget asked children 70 years ago. Pour water into a taller, thinner glass: is it the same amount? Seven-year-olds pass this reliably. We built ConservationBench of 384 real videos of conservation tasks with matched non-conserving controls, and evaluated 112 VLMs across 23,040 trials. Humans score 98 percent; models sit near chance, with 82 of 112 under 10 percent on the strict paired test. The most telling result: models answer "better with no image at all" than with the actual video, and chain-of-thought prompting makes things worse, showing they're leaning on a textual prior, not tracking object state through change.

#videomodel #AI #research #reasoning

[Yaoyao(Freax) Qian @RubyFreax]

Just open-sourced TermHub!😍

A terminal-style academic homepage template built for AI workflows.

Turn your CV into Markdown with ChatGPT / Claude, plug it into the template, and deploy.

Guide
https://t.co/ZHHSXFsBAw
Repo
https://t.co/FGRVT8J4PR

[Hokin Deng @DengHokin]

Yes! We are top paper of the week 🚀🚀🚀

[Yaoyao(Freax) Qian @RubyFreax]

@DengHokin 👏Amazing leadership on this project!

[Yaoyao(Freax) Qian @RubyFreax]

Glad to have been part of this! There's genuinely a lot of interesting stuff in the data and tools if you're into video reasoning https://t.co/eiv6D8TU6u

[Hokin Deng @DengHokin]

#VideoReason We are open-sourcing the entire VBVR stack to speed-up the arrival of video reasoning as the next fundamental paradigm of intelligence

- 150+ synthetic generators
- 1 million training clips
- Cloud-scale data factory
- Unified EvalKit
- 100 rule-based evaluators
- Strong baseline model

Checkout at https://t.co/lOtJzJYC52

[Yaoyao(Freax) Qian @RubyFreax]

@SonglinYang4 @MITEECS @thinkymachines 🤩🤩🤩Congrats!!!!!

[Yaoyao(Freax) Qian @RubyFreax]

Off to San Diego for #NeurIPS2025 ! 🌴
I’ll be there the whole week (Dec 2–7).
If you’re around and want to talk research, I’m always up for it.
DM me if you want to meet up!

[Andrej Karpathy @karpathy]

I quite like the new DeepSeek-OCR paper. It's a good OCR model (maybe a bit worse than dots), and yes data collection etc., but anyway it doesn't matter.

The more interesting part for me (esp as a computer vision at heart who is temporarily masquerading as a natural language person) is whether pixels are better inputs to LLMs than text. Whether text tokens are wasteful and just terrible, at the input.

Maybe it makes more sense that all inputs to LLMs should only ever be images. Even if you happen to have pure text input, maybe you'd prefer to render it and then feed that in:
- more information compression (see paper) => shorter context windows, more efficiency
- significantly more general information stream => not just text, but e.g. bold text, colored text, arbitrary images.
- input can now be processed with bidirectional attention easily and as default, not autoregressive attention - a lot more powerful.
- delete the tokenizer (at the input)!! I already ranted about how much I dislike the tokenizer. Tokenizers are ugly, separate, not end-to-end stage. It "imports" all the ugliness of Unicode, byte encodings, it inherits a lot of historical baggage, security/jailbreak risk (e.g. continuation bytes). It makes two characters that look identical to the eye look as two completely different tokens internally in the network. A smiling emoji looks like a weird token, not an... actual smiling face, pixels and all, and all the transfer learning that brings along. The tokenizer must go.

OCR is just one of many useful vision -> text tasks. And text -> text tasks can be made to be vision ->text tasks. Not vice versa.

So many the User message is images, but the decoder (the Assistant response) remains text. It's a lot less obvious how to output pixels realistically... or if you'd want to.

Now I have to also fight the urge to side quest an image-input-only version of nanochat...

[Yinpei Dai @YinpeiD]

Thanks @_akhaliq for sharing our work!
Aim and Grasp! AimBot introduces a new design to leverage visual cues for robots - similar to scope reticles in shooting games.
Let's equip your VLA models with low-cost visual augmentation for better manipulation!
https://t.co/bSXPzml51B

[Guohao Li 🐫 @guohao_li]

Introducing Eigent — the first multi-agent workforce on your desktop.

Eigent is a team of AI agents collaborating to complete complex tasks in parallel. It is your long-term working partner with fullly customizable workers and MCPs.

Public beta available to download for MacOS, Windows. 100% open-source on Github. Comment for 500 extra credits.

[Haibo Zhao @ZhaoHaibo47588]

Excited to share our #ICML2025 paper, Hierarchical Equivariant Policy via Frame Transfer. Our Frame Transfer interface imposes high-level decision as a coordinate frame change in the low-level, boosting sim performance by 20%+ and enabling complex manipulation with 30 demos.

[Yaoyao(Freax) Qian @RubyFreax]

Owen will be presenting our poster for the paper Hierarchical Equivariant Policy via Frame Transfer at ICML Today (see https://t.co/WAevLYDzyO for details). If you are interested in equivariance and/or robotic manipulation please stop by!

[Yaoyao(Freax) Qian @RubyFreax]

inspired talk! #RSS2025

[Yaoyao(Freax) Qian @RubyFreax]

🤣First time at RSS!

Happy to meet up and chat!

[Infini-AI-Lab @InfiniAILab]

🔥 We introduce Multiverse, a new generative modeling framework for adaptive and lossless parallel generation.

🚀 Multiverse is the first open-source non-AR model to achieve AIME24 and AIME25 scores of 54% and 46%

🌐 Website: https://t.co/J9osByhWUf

🧵 1/n

[Songlin Yang @SonglinYang4]

📢 (1/16) Introducing PaTH 🛣️ — a RoPE-free contextualized position encoding scheme, built for stronger state tracking, better extrapolation, and hardware-efficient training. PaTH outperforms RoPE across short and long language modeling benchmarks
https://t.co/nJItUuYKWZ

[Yaoyao(Freax) Qian @RubyFreax]

Proud to be part of this open-source effort after joining PathOnAI! 🌱

We hope this helps push web agent research toward more robust, interpretable, and deployable systems.

[Yaoyao(Freax) Qian @RubyFreax]

🚀 Excited to share VisualTreeSearch, my first project + upcoming paper with the open-source research group @PathOnAI !

It is a fully-deployed system for understanding test-time tree search in web agents, now open-source & demo-ready. 👇

[Yaoyao(Freax) Qian @RubyFreax]

📄 Paper coming soon.
🎥 Video + docs in progress.
🧪 Explore the system now:
Project: https://t.co/ZhSu0sYvto
Demo: https://t.co/CB4CP5SKIc
Code: https://t.co/S6F8aXoFMa