Junyi Zhang @ CoRL

Ph.D. @Berkeley_AI.
research intern @physical_int.
ex @GoogleDeepMind, @MSFTResearch.
vision, generative model, robotics.

[Junyi Zhang @ CoRL @junyi42]

𝗢𝗻𝗲 𝗺𝗲𝗺𝗼𝗿𝘆 𝗰𝗮𝗻’𝘁 𝗿𝘂𝗹𝗲 𝘁𝗵𝗲𝗺 𝗮𝗹𝗹.

We present 𝗟𝗼𝗚𝗲𝗥, a new 𝗵𝘆𝗯𝗿𝗶𝗱 𝗺𝗲𝗺𝗼𝗿𝘆 architecture for long-context geometric reconstruction.

LoGeR enables stable reconstruction over up to 𝟭𝟬𝗸 𝗳𝗿𝗮𝗺𝗲𝘀 / 𝗸𝗶𝗹𝗼𝗺𝗲𝘁𝗲𝗿 𝘀𝗰𝗮𝗹𝗲, with 𝗹𝗶𝗻𝗲𝗮𝗿-𝘁𝗶𝗺𝗲 𝘀𝗰𝗮𝗹𝗶𝗻𝗴 in sequence length, 𝗳𝘂𝗹𝗹𝘆 𝗳𝗲𝗲𝗱𝗳𝗼𝗿𝘄𝗮𝗿𝗱 inference, and 𝗻𝗼 𝗽𝗼𝘀𝘁-𝗼𝗽𝘁𝗶𝗺𝗶𝘇𝗮𝘁𝗶𝗼𝗻.

Yet it matches or surpasses strong optimization-based pipelines. (1/5)

@GoogleDeepMind @Berkeley_AI

[Junyi Zhang @ CoRL @junyi42]

@redstone_hong Congrats Hongsuk!!

[Junyi Zhang @ CoRL @junyi42]

We will be presenting our poster at 10:30am at JW Marriott, 3rd floor, Poster #99. Come chat with us about agentic robotics! 👀

[Jiaxin Ge @ CoRL @jiaxin_ge_]

@junyi42 and I will be giving our spotlight talk on Playful Agentic Robot Learning (RATs) in CoRL at 8:30 tomorrow! Come checkout our paper😃

[Jiaxin Ge @ CoRL @jiaxin_ge_]

@junyi42 and I will be giving our spotlight talk on Playful Agentic Robot Learning (RATs) in CoRL at 8:30 tomorrow! Come checkout our paper😃

[Yao Tang @tyao923]

When an agent meets a program: 𝗣𝗹𝗮𝘆 𝗶𝘁. 𝗥𝗲𝗯𝘂𝗶𝗹𝗱 𝗶𝘁.

𝗣𝗿𝗼𝗴𝗿𝗮𝗺𝗟𝗮𝗯 gives agents 1,261 real-world programs to practice on.

𝗧𝗿𝗮𝗶𝗻 𝗼𝗻𝗲 𝗮𝗴𝗲𝗻𝘁, 𝗼𝗿 𝗮 𝘄𝗵𝗼𝗹𝗲 𝘁𝗲𝗮𝗺.

🧵0/n

[Junyi Zhang @ CoRL @junyi42]

We trained an agent with native visual world modeling ability.

It imagines counterfactual outcomes to plan ahead. SFT teaches it to imagine; agentic RL makes it even more adaptable.

We’ve released everything: https://t.co/ERQYbmHQUg

[Jiaxin Ge @ CoRL @jiaxin_ge_]

We trained an imaginative agent! ✨
Agima: an omni agent that plans by imagining futures.

𝗦𝗙𝗧: Directly generalizes to harder tasks! 𝟮%→𝟰𝟬%, beats GPT-5.6 Sol (𝟮𝟬%).
𝗥𝗟: New ways to imagine emerge! Imagines a second plan only if the first one fails (𝟯𝟮%→𝟭𝟬𝟬%).

[Donghoon Ahn @donghoon_ahn]

Can diffusion models generate high-quality images without CFG?

Yes! We introduce ✨Correction Training for Diffusion Models✨, which significantly improves sample quality without relying on classifier-free guidance.

Project page: https://t.co/Ic8foxITJ0

🧵 (1/n)

[Haven Feng ✈️ CoRL @HavenFeng]

Now autoregressive video models can remember. 🧠

Introducing MosaiChunk: persistent visual memory for long-horizon autoregressive video generation🧩.

Instead of remembering everything, it remembers the right things. ✍️

1/6

[Jiaxin Ge @ CoRL @jiaxin_ge_]

Generation 𝗖𝗔𝗡 improve understanding!

With the right recipe, we show that training on generation data can help understanding tasks.

Introducing OmniTaskonomy: our year-long study of when and how generation helps understanding.

[Zihan Wang @Z1hanW]

Humans generalize from rich interaction experience. We bring this to humanoid loco-manipulation through scalable real-to-sim-to-real.

Introducing PRISM, we expand 4 real videos into 256 counterfactual variants, recover robot-object trajectories, and train one policy that generalizes across diverse objects, scales, and spatial layouts.

CoRL’26 🚀 Website for details 👉 https://t.co/DUf2DIrvai

Work done at Amazon FAR w/ @zhenkirito123 @pabbeel @rocky_duan @JitendraMalikCV @carlo_sferrazza @ckarenliu @GuanyaShi @akanazawa

[Jiaxin Ge @ CoRL @jiaxin_ge_]

Our paper RATs has been selected as CoRL spotlight!

We apply "sleep time scaling" to agentic robotics learning. The agent aquires skills during play and then uses these skills at test time. Check out our project page at : https://t.co/xnZcBRgJPV @junyi42

[Jiaxin Ge @ CoRL @jiaxin_ge_]

Our paper VisGym has been accepted to NeurIPS 2026!

It’s been crazy to see how the field has developed since we have released this paper. Hope that VisGym can continue to serve as a diagnostic playground to identify failure pattern for new frontier multimodal agents, and serve for the development of this field.

[Wenli Xiao 🚀CoRL @_wenlixiao]

I didn't get what GPT-6 meant for robotics until I actually tried it.

GPT-6 Astra just does physical ICL out of the box.

we drop a recording of a human doing a novel task into 𝗰𝗼𝗱𝗲𝘅 app. Prompt it to drive a robot arm the same way.

It just works on the first pass!

[Haocheng Xi @HaochengXiUCB]

Open-source video generation is now faster than playback without compromising quality.

Introducing Video Delta Net (VDN): hybrid attention for live text-to-video with near-lossless quality.

VDN accelerates Minimax-H3 by 75 - 90 x, generating 14 seconds of 768p video in 11 seconds on 8× NVIDIA B200 GPUs.

Checkpoints + training/inference code + Technical Blog ⬇️
(1/6)

[Junyi Zhang @ CoRL @junyi42]

@Stone_Tao @physical_int Congrats Stone!!

[Junyi Zhang @ CoRL @junyi42]

"The latent world representation is a program, not a vector"

Finding the right representation to be inverse-graphics is really important. Really cool work!

[Haven Feng ✈️ CoRL @HavenFeng]

Today, we’re introducing [schema]: a harness reaching 99% RHAE with Opus 4.8 + Fable 5 and 95.35% with GPT-5.6 Sol on ARC-AGI-3 Public set.

[schema] makes an LLM think like a physicist. 🧵

[Martin Ziqiao Ma @ziqiao_ma]

The term "continual learning" has become overloaded if you see it as an ML problem.

One classic thread is about memorization: regularization-based continual learning methods, such as EWC, MAS, and SI, estimate which parameters mattered for previous tasks and resist changing them too much.

One modern thread is about adaptation: test-time training and inference-time learning methods, such as TTT, adapt part of the model on the incoming test stream before making predictions.

These are sometimes discussed as separate threads. But in modern scalable architectures, I think they are better seen as complementary constraints: a model that learns quickly at test time also benefits from a mechanism for deciding what not to forget.

In our #ECCV2026 paper, we study this in large-scale 4D reconstruction: how to build fast spatial memory that can adapt over long observation streams while reducing collapse and forgetting.

Instead of using fully plastic test-time updates, we stabilize fast-weight adaptation with an elastic prior that balances adaptation and memory.

Key ideas:
- Elastic Test-Time Training: Fisher-weighted consolidation for fast-weight updates
- EMA anchor weights that provide a moving reference for stability
- Chunk-by-chunk inference for long 3D/4D observation streams

We show that this scales across large 3D/4D pretraining settings, including both LRM-style and LVSM-style models, and improves reconstruction across benchmarks including Stereo4D, NVIDIA, and DL3DV-140.

We release model checkpoints across different design choices: resolution, post-training curriculum, and whether the model uses an explicit 4DGS intermediate representation.
- Homepage: https://t.co/7iXjTiKcdu
- Paper: https://t.co/SBUu9kGS87
- Code: https://t.co/FMrStj4fW6
- Models: https://t.co/8lo3Dk95en

This work is co-led with @Xueyang_Y, contributed by @zhnhoy5 @YuncongYY, and advised by @SLED_AI @gan_chuang.

[Ken Goldberg @Ken_Goldberg]

Great work using offline agentic exploration to develop robot skills!

[Baifeng @baifeng_shi]

Learning from task-agnostic, explorative experience!

[Dantong Niu @Dantong_Niu]

Excited to share T-Rex: Tactile-Reactive Dexterous Manipulation 🦖🤖

Touch is fundamental to human dexterity, yet most Vision-Language-Action (VLA) models either ignore tactile feedback or lack the ability to react to high-frequency contact signals.

In this work, we tackle both the data and architectural challenges of tactile-reactive dexterous manipulation.

🦖 A 100-hour tactile-synchronized dexterous manipulation dataset with 7,700+ trajectories, 22 motor primitives, and 200+ everyday objects.

🦖 A tactile-reactive MoT architecture with spatial-temporal tactile encoding and asynchronous high-frequency tactile refinement.

🦖 A scalable training recipe combining 22,889 hours of human egocentric pretraining with tactile-grounded robot mid-training.

Across 12 real-world contact-rich manipulation tasks, T-Rex achieves over 30% higher average success rate than the strongest baseline.

We are fully open-sourcing the dataset, models, teleoperation stack, training code, and inference pipeline.

🌐 Project: https://t.co/AiHKRR8YXU
📄 Paper: https://t.co/mXY2UNLlqc
💻 Code: https://t.co/7skCxUtwKC
🤗 Dataset: https://t.co/uNwW8dcRZL
🧵 Thread ↓

[Junyi Zhang @ CoRL @junyi42]

That's a nice question! Currently, we threshold the play budget so that it’s comparable or less to the amount of compute used at test time. We also include a cost analysis in the appendix. That said, I think smarter ways of controlling the budget are definitely an interesting direction!

[Junyi Zhang @ CoRL @junyi42]

Children learn from play. Can robots do the same?

We propose 𝐏𝐥𝐚𝐲𝐟𝐮𝐥 𝐀𝐠𝐞𝐧𝐭𝐢𝐜 𝐑𝐨𝐛𝐨𝐭 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠, a paradigm that gives embodied coding agents a play stage before downstream tasks arrive, and instantiate it with 𝐑𝐀𝐓𝐬 (Robotics Agent Teams), where robots discover reusable skills through curious play.

Co-led with @jiaxin_ge_