cogtoolslab/physics-benchmarking-neurips2021

Repo for "Physion: Evaluating Physical Prediction from Vision in Humans and Machines", presented at NeurIPS 2021 (Datasets & Benchmarks track)

33
/ 100
Emerging

This project provides a comprehensive dataset and tools for evaluating how well AI models (and humans) can predict physical outcomes from visual information. It takes video clips of objects interacting under various physical scenarios (like colliding or dropping) as input. The output is a performance comparison, showing how accurately models and humans predict physical events. This is ideal for researchers in AI, cognitive science, and robotics working on visual perception and physics-based reasoning.

No commits in the last 6 months.

Use this if you need high-quality, pre-verified video datasets and evaluation metrics to benchmark AI models against human performance in tasks requiring physical prediction from vision.

Not ideal if you are looking for a general-purpose physics simulation engine or a tool for real-time physical inference in uncontrolled environments.

AI-benchmarking cognitive-modeling physics-based-vision machine-perception robotics-research
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 8 / 25

How are scores calculated?

Stars

85

Forks

5

Language

Jupyter Notebook

License

MIT

Last pushed

Feb 09, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/cogtoolslab/physics-benchmarking-neurips2021"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.