Twitxy Saved

Ayush Tewari

@_atewari

Assistant Professor @Cambridge_Eng
ayushtewari.com Joined December 2018
639 Following 2.1K Followers 134 Posts

Latest posts 5

Pinned post

Ayush Tewari @_atewari
Video models can be physically and logically consistent! Check out S2PD - we show consistent generations on a large set of games and simulations. Bidirectional models fail to generate consistent outputs on most of these datasets.

jefequien.github.io/S2PD/
Jeffrey Hu Verified account @jefequien Oct 6
Introducing S2PD: Serial-to-Parallel Diffusion.

Bidirectional video models fail to follow simple rules even when trained on enormous amounts of data. We show that serial computation is critical for generating consistent video and use it in the high-noise regime of video diffusion.

Project Page: jefequien.github.io/S2PD/
View quoted post on X
1 replies 11 reposts 113 likes 6.1K views View on X
Ayush Tewari @_atewari
Check out 3D-Belief: a 3D world model that explores how explicit belief representations can support memory, imagination, and planning under partial observability.
Yifan Yin @yifanyin_11 Jun 1
What should a world model capture for embodied agents?

An agent acting under partial observability needs a belief over the 3D world: what it has seen, what may exist beyond view, and how that belief should update as it moves.

Introducing 3D-Belief, a 3D world model for embodied belief inference under partial observability. 🧵[1/9]
View quoted post on X
Ayush Tewari @_atewari
Is pixel prediction the best way to build a world model?

Check out VDAWorld, an alternative path to building interpretable, editable, and physically grounded world models.

We use a VLM to build a simulation of the scene with the help of a computer vision toolbox.
0:05
6 replies 14 reposts 165 likes 9.4K views View on X