ECCV 2026 Workshop · Full-day · In-person

Physical AI:
Understanding and Building the
Physical World

The 1st Workshop on Physical AI brings together researchers across vision, graphics, robotics, and generative modeling to advance physics-grounded understanding of the real world — from physical property estimation and 3D/4D reconstruction to differentiable simulation and physically plausible generation.

Venue ECCV 2026 · Malmö, Sweden
Date, Location September 9, 2026 · Malmömässan A
Format 10 Talks · 4D Challenge
01 · Overview

About the workshop

Physical AI seeks to endow AI systems with a deep, physics-grounded understanding of the real world. While recent advances in computer vision have enabled large-scale geometric reconstruction, multimodal reasoning, and generative modeling, most systems remain limited in modeling physical properties — mass, friction, material behavior, structural stability, deformation, and dynamic interactions.

This workshop integrates physical reasoning throughout the pipeline: physical property estimation, physics-informed 3D/4D reconstruction, differentiable simulation, physically plausible generation, and embodied interaction. Rather than treating physics as a downstream refinement, PhysAI positions it as a core inductive bias for representation learning and world modeling.

The workshop fosters interdisciplinary discussion across vision, graphics, robotics, and digital twinning, and is built around 10 invited talks and a competition on dynamic 4D reconstruction.

02 · Scope

Topics of interest

01

Physical scene understanding

Physical attribute estimation and reasoning — materials, mass, friction, affordances.

02

Physics-aware 3D/4D reconstruction

Reconstruction from sparse or unconstrained inputs that respects physical priors.

03

Generative world models & world simulators

Generative models and world simulators for physically consistent world modeling — a key recent goal of physical-world AI.

04

Spatial & physical reasoning

Reasoning from images, videos, and multi-modal data about physical behavior.

05

Embodied AI & robot learning

Robot learning in physics-grounded virtual environments and digital twins.

06

Benchmarks & datasets

New benchmarks and datasets for physical scene understanding and reconstruction.

03 · Invited speakers

Confirmed & invited speakers

A line-up of leaders shaping the next generation of physics-grounded AI. Listed alphabetically.

* Speaker list is tentative; confirmations to be announced.

04 · Schedule

Workshop program

Full-day workshop · 10 invited talks · competition highlights and awards.

Time
Session
08:50 – 09:00
Opening remarks — Andrea Vedaldi
09:00 – 09:35
Invited Talk 1 — Andreas Geiger
09:35 – 10:10
Invited Talk 2 — Yogesh Balaji
10:10 – 10:25
☕ Coffee break
10:25 – 11:00
Invited Talk 3 — Vincent Sitzmann
11:00 – 11:35
Invited Talk 4 — Jiajun Wu
11:35 – 12:10
Invited Talk 5 — Kristen Grauman
12:10 – 14:00
🥗 Lunch break
14:00 – 14:35
Invited Talk 6 — Mike Roberts
14:35 – 15:10
Invited Talk 7 — Matthias Nießner
15:10 – 15:45
Invited Talk 8 — David Novotny
15:45 – 16:20
Invited Talk 9 — Marc Pollefeys
16:20 – 16:55
Invited Talk 10 — Haiwen (Haven) Feng
16:55 – 17:30
Competition highlights — Awards & winning solutions
17:30 – 17:40
Closing remarks

Times are local to the ECCV 2026 venue and are subject to minor changes.

05 · Competition

PhysAI Dynamic 4D Reconstruction Challenge

A two-track benchmark for reconstructing the physical world in motion.

Test submission deadline August 10, 2026 · 23:59 UTC
Explore the challenge →
06 · Organizers

Organizing committee

A team spanning vision, graphics, robotics, and generative modeling — across Oxford VGG, NTU PVG, NTU MMLab, ETH Zurich, Google DeepMind, NAVER LABS Europe, and Ropedia.

07 · Sponsor

Our sponsor

We gratefully acknowledge Ropedia's support of PhysAI 2026 and the USD 2,000 award for the PhysAI Dynamic 4D Reconstruction Challenge winner.

08 · Contact

Get in touch

For questions about the workshop, the challenge, or sponsorship opportunities, please reach out:

Hosted at ECCV 2026, Malmö, Sweden
Room Malmömässan A
Follow updates Twitter/X · LinkedIn