# Learning A Physical Long-term Predictor

arXiv: https://arxiv.org/abs/1703.00247 · HTML: https://mlsanity.com/paper/1703.00247/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Mar 2017.

## Built on by (most cited first, 5 of 5)

- [gradSim: Differentiable simulation for system identification and visuomotor control](https://mlsanity.com/paper/2104.02646/) (2021)
- [Inferring Fluid Dynamics via Inverse Rendering](https://mlsanity.com/paper/2304.04446/) (2023)
- [Distilling Governing Laws and Source Input for Dynamical Systems from Videos](https://mlsanity.com/paper/2205.01314/) (2022)
- [Uncovering Closed-form Governing Equations of Nonlinear Dynamics from Videos](https://mlsanity.com/paper/2106.04776/) (2021)
- [RoboGolf: Mastering Real-World Minigolf with a Reflective Multi-Modality Vision-Language Model](https://mlsanity.com/paper/2406.10157/) (2024)

## Abstract (arXiv)

Evolution has resulted in highly developed abilities in many natural intelligences to quickly and accurately predict mechanical phenomena. Humans have successfully developed laws of physics to abstract and model such mechanical phenomena. In the context of artificial intelligence, a recent line of work has focused on estimating physical parameters based on sensory data and use them in physical simulators to make long-term predictions. In contrast, we investigate the effectiveness of a single neural network for end-to-end long-term prediction of mechanical phenomena. Based on extensive evaluation, we demonstrate that such networks can outperform alternate approaches having even access to ground-truth physical simulators, especially when some physical parameters are unobserved or not known a-priori. Further, our network outputs a distribution of outcomes to capture the inherent uncertainty in the data. Our approach demonstrates for the first time the possibility of making actionable long-term predictions from sensor data without requiring to explicitly model the underlying physical laws.