Anujraaj Goyal ✈️ NeurIPS2025

current @snap //prev @carnegiemellon

[Bharath Raj @bharathrajn98]

Feed-forward 3D reconstruction methods typically predict pointmaps in camera-centric frames. But why should a camera's arbitrary orientation define the coordinate system?

We introduce G3T, a transformer that predicts pointmaps in gravity-aligned frames. Regardless of input image orientation, our method always produces upright pointmaps (see demo).

We leverage this uprightness to create G3T-Long, a submap-based reconstruction method that improves robustness on long-sequence 3D reconstruction (more on that below).

Interactive demos, code, and model weights are available on our project page.

[Gordon (Guocheng) Qian @guocheng_qian]

Excited to share that EasyV2V has been accepted to CVPR 2026! Data is the fuel. EasyV2V introduces a comprehensive and scalable video editing data pipeline. Curious how different types of data impact video editing performance? Check out our project page:
https://t.co/IsQkUoiVE2

[Gordon (Guocheng) Qian @guocheng_qian]

🚀New paper Omni-Attribute is released!
This work can isolate a specific attribute from any image and merge those selected attributes from multiple images into a coherent generation.
It’s a pleasure working with @tsaishien_chen, the amazing team at @Snap, and @junyanz89 from CMU!

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@guocheng_qian @kcjacksonwang @AbermanKfir @c414driu5 @vehsatso @SergeyTulyakov Huge thanks to my brilliant collaborators at Snap who contributed through countless experiments, failures, redesigns, and aha moments. Couldn’t have done this without them.

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

I’m excited to be at NeurIPS 2025, presenting our latest work “Preventing Shortcuts in Adapter Training via Providing the Shortcuts”

📅 Wed, Dec 3, 2025
⏰ 11:00 AM – 2:00 PM PST
📍 Exhibit Hall C, D, E — Poster #4406

📄 arXiv: https://t.co/At21XrLGmb

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@guocheng_qian @kcjacksonwang @AbermanKfir @c414driu5 @vehsatso @SergeyTulyakov This became the key insight behind this work

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@guocheng_qian @kcjacksonwang @AbermanKfir @c414driu5 @vehsatso @SergeyTulyakov We found something surprisingly simple and counterintuitive: If you don’t want your model to learn a particular attribute, explicitly provide that attribute during training — and it stops trying to learn it!

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@guocheng_qian @kcjacksonwang @AbermanKfir @c414driu5 @vehsatso @SergeyTulyakov Why do adapters copy-paste extra garbage from the input image — background, lighting, pose, expression — instead of just learning subject’s identity?

[Snap Inc. @Snap]

@CVPR Here's a full look at our programming. ⬇️

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

Happy to chat about image/video personalisation, on-device diffusion models, and life beyond work!

Looking fwd to connecting with old friends and new at #CVPR2025

[Snap Inc. @Snap]

Heading to @CVPR 2025 in Nashville this week? So are we!

We’re proud to have 12 papers accepted — including SnapGen and 4Real-Video, both highlighted among the top 3% of submissions.

Come find us to learn more about the cutting edge work we’re doing in AI and computer vision.

📍 See you in Nashville!

Learn more: https://t.co/JQIghmmgo6

[Gordon (Guocheng) Qian @guocheng_qian]

📢I am attending #CVPR2025 (Jun 11 - 14). Come to our https://t.co/vAx6y3opa9 poster to know more about how we achieved the highest ID preservation in personalization and further enables expression following in our follow ups. See you at Fri 4 - 6 pm, ExHall D Poster #326.

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

Check out our video explanation: https://t.co/0khnRwyNOy

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

Can we predict 3D human posture and body pressure from a single depth and pressure image?

Check out our #CVPR2024 work, BodyMAP: Jointly Predicting Body Mesh & 3D Applied Pressure Map for People in Bed 🧵

📅 Poster: Wed 19 Jun, 10:30 a.m. Arch 4A-E #222 https://t.co/PJn29fvRi7

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

Paper: https://t.co/fsRkGO6hV0
Code: https://t.co/V0so31eZz0
Project Page: https://t.co/PJn29fvRi7

w/ Abhishek Tandon @henrymclever @ZackoryErickson

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@MIT_CSAIL describe_coding(n-1) + “\n” ;

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@kritinotkeerti Now imagine they crash the wedding? That won’t be you, don’t worry :)

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@samridhi_m Talking about hiding in plain sight. Thx!

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@samridhi_m Dayummm!

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@dalmiaman Are you back in town?

[Anujraaj Goyal ✈️ NeurIPS2025 @timbektu]

@kkitani What questions should I ask myself to figure if I will enjoy/make a better engineer or a researcher?