Anurag Bagchi

CMU, ex-TikTok AI

[Anurag Bagchi @Miccooper9]

[1/6] Ego-centric World Models

We introduce EgoWM — a video world model that simulates EVE-1X humanoid interactions from a single ego-view image + full-body joint angle trajectories.

Moreover it effortlessly generalizes to extreme OOD domains, including paintings !

[Aryan Satpathy @satpathyaryan45]

Excited to share our project - Sim2Reason!

Key Insight: Simulators are an untapped source of cheap supervision for scientific reasoning. LLMs can learn physical reasoning from simulation to improve on real world benchmarks such as the International Physics Olympiad!

[Vincent Sitzmann @vincesitzmann]

In my recent blog post, I argue that "vision" is only well-defined as part of perception-action loops, and that the conventional view of computer vision - mapping imagery to intermediate representations (3D, flow, segmentation...) is about to go away.

https://t.co/aFmE9CHHau

[Anurag Bagchi @Miccooper9]

@zhihelu1 Thanks! for world models, more precisely, forward dynamics models (current state + action -> future state), this is standard formulation. There are lots of model-based control approaches that can be used to plan/predict actions using such world models.

[Anurag Bagchi @Miccooper9]

[1/6] Ego-centric World Models

We introduce EgoWM — a video world model that simulates EVE-1X humanoid interactions from a single ego-view image + full-body joint angle trajectories.

Moreover it effortlessly generalizes to extreme OOD domains, including paintings !

[Anurag Bagchi @Miccooper9]

[6/6] Fine-grained humanoid manipulation

EgoWM enables precise 25-DoF joint-angle manipulation with the EVE-1X humanoid,
even at 4× temporal compression (Cosmos-2B).

📷 Learn more:
Project page: https://t.co/GgyMU0sF9v
Paper: https://t.co/2LNOIi6MCs

[Anurag Bagchi @Miccooper9]

[5/6] Temporal compression

Unlike prior works, we preserve full-sequence diffusion and compress actions to the latent temporal resolution. EgoWM achieves +42% better action alignment at +4s horizon vs. frame-wise autoregressive NWMs
even with 4× temporal compression (Cosmos-2B).

[Shraman Pramanick @Shramanpramani2]

My role at Meta's SAM team (MSL, previously at FAIR Perception) has been impacted within 3 months of joining after PhD.

If you work with multimodal LLMs for grounding or complex reasoning, or have a long-term vision of unified understanding and generation, let's talk.

I am on the job market starting immediately.

#metalayoffs #FAIR #MSL #SAM

[Anurag Bagchi @Miccooper9]

Happening now! ICCV 25 poster#323
Drop by to chat and see some cool results!

[Anurag Bagchi @Miccooper9]

[ICCV 25] Refer Everything Model (REM)

(1/6) We leverage Text-to-Video Generation models to zero-shot segment any concept in a video using text. REM generalises to dynamic concepts like smoke, light-beam and more without ever having seen segmentation masks for these entities.

[Anurag Bagchi @Miccooper9]

@andrew_n_carr Thanks Andrew! We were also really surprised to see how well this worked. Exciting times ahead.

[Anurag Bagchi @Miccooper9]

[ICCV 25] Refer Everything Model (REM)

(1/6) We leverage Text-to-Video Generation models to zero-shot segment any concept in a video using text. REM generalises to dynamic concepts like smoke, light-beam and more without ever having seen segmentation masks for these entities.

[Anurag Bagchi @Miccooper9]

(6/6) We’re at the start of the internet-scale "video" era, and the possibilities are exciting. Learn more at https://t.co/ERyk8Cfpst — our code & model weights are available. Visiting ICCV? Come see our poster on Oct 23 to chat and see results in action!

[Anurag Bagchi @Miccooper9]

(5/6) REM demonstrates how Text-to-Video generation can serve as a powerful pre-training paradigm for downstream video understanding. The days of large-scale, labor-intensive video annotation may soon be behind us — pre-train to generate, fine-tune lightly to understand.