Projects & Demo

Projects & Demo

COM4D

COM4D

Compositional 4D scene reconstruction from monocular video — CVPR 2026.

Jun 15, 2026

FireScope

FireScope

Wildfire risk mapping with a chain-of-thought oracle — CVPR 2026.

May 29, 2026

EgoNight

EgoNight

Benchmark for egocentric vision understanding at night — ICLR 2026.

Apr 21, 2026

VOID

VOID

Video object and interaction deletion — INSAIT × Netflix.

Apr 8, 2026

AR-VLA

AR-VLA

Autoregressive action expert for vision–language–action models — RSS 2026.

Mar 10, 2026

SPEAR-1

SPEAR-1

Scaling beyond robot demonstrations via 3D understanding — open weights from INSAIT Robotics.

Mar 1, 2026

Agentic AI for Visual Media Workshop

Agentic AI for Visual Media Workshop

1st Workshop on Agentic AI for Visual Media — Denver, June 2026.

Feb 18, 2026

DiffSim Trinity

DiffSim Trinity

End-to-end autonomous driving via differentiable simulation.

Jan 12, 2026

GaussianVLM

GaussianVLM

Scene-centric 3D VLM on language-aligned Gaussian splats — RA-L 2025.

Dec 1, 2025

MotoVLA

MotoVLA

Generalist robot manipulation beyond action-labeled data — CoRL 2025.

Nov 6, 2025

SceneSplat

SceneSplat

Open-vocabulary 3DGS scene understanding with vision-language pretraining — ICCV 2025 Oral.

Oct 13, 2025

ObjectRelator

ObjectRelator

Cross-view ego–exo object correspondence with MCFuse and XObjAlign — ICCV 2025 Highlight.

Oct 13, 2025

GAIA Symposium

GAIA Symposium

Geospatial AI and Applications with Foundation Models — Sofia, September 2025.

Sep 24, 2025