Angjoo Kanazawa

@akanazawa · Berkeley, CA · joined 04 Jun 2011

Assistant Professor at @Berkeley_EECS, @berkeley_ai. KAIR, @nerfstudioteam. Amazon Scholar @ FAR. Previously advised @WonderDynamics and @LumaLabsAI. she/her.

21 799Followers

623Following

855Posts total

354.4KViews on collected posts

Son gönderiler

Angjoo Kanazawa@akanazawa16 Jul

We got 99% on ARC-AGI-3 by instilling analysis-by-synthesis into AI models. [Schema] is basically VIGA squared: do inverse graphics, then do it again one level up — inverse dynamics. Super simple, surprisingly effective! Analysis by synthesis FTW!!! 35.7K views · 213 likes · 13 reposts · 10 replies Open on X →

Angjoo Kanazawa@akanazawa17 Jun

So excited for this to finally be out: A fully open-source data & stack released so we can learn the ABCs of bimanual manipulation together! This was a huge team effort, and many long days and nights of work. Congrats to the amazing team, I'm very proud of you! https://t.co/Zv

0:45

29.5K views · 182 likes · 21 reposts · 4 replies Open on X →

Angjoo Kanazawa@akanazawa22 May

Babies learn by being naturally curious. How do we get autonomous agents to do the same? We revisited curiosity in 3D exploration and found that memory is key. This project taught me a lot about what kind of functions an agent and a "world model" need to have for this direction 34.1K views · 148 likes · 20 reposts · 2 replies Open on X →

Angjoo Kanazawa@akanazawa22 May

🚀 🚀 🚀 Excited to share our new paper: Remember to be Curious: Episodic Context and Persistent Worlds for 3D Exploration What does it take for an agent to stay curious in a 3D world? The answer is memory. 🌐 Project: https://t.co/Nw9eWb2dCB 📄 Paper: https://t.co/78GE9CZE15 https

1:12

72.2K views · 223 likes · 41 reposts · 2 replies Open on X →

Angjoo Kanazawa@akanazawa21 Apr

Very excited to share this work @davidrmcall did with the fantastic NVIDIA Finland team last year. We have a surprisingly simple, but sample efficient way to post-train a flow model with RL. 20.6K views · 109 likes · 12 reposts · 3 replies Open on X →

Angjoo Kanazawa@akanazawa17 Apr

We developed a simple, sample-efficient online RL technique for post-training image generation models. We see it as a possible steerable alternative to CFG, driven by any scalar reward, including human preference. https://t.co/41CzipIQKR

0:26

67.7K views · 376 likes · 42 reposts · 11 replies Open on X →

Angjoo Kanazawa@akanazawa31 Mar

Diffusion forcing is great for sequence modeling! We've been working on social behavior prediction but nothing's worked this well. Representation matters: We encode motion with discrete latents & model relative relationships in a neat way. It handles multiple tasks & people! 27.3K views · 153 likes · 16 reposts · 3 replies Open on X →

Angjoo Kanazawa@akanazawa27 Mar

When people share a space, their movements become intertwined. Embodied agents need to understand these social dynamics to interact effectively. Introducing MAGNet 🧲, a unified autoregressive diffusion forcing model for multi-agent motion generation that captures these https://t

0:38

67.3K views · 371 likes · 58 reposts · 8 replies Open on X →

Growth & engagement

How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.