# Physical Reasoning Using Dynamics-Aware Models

arXiv: https://arxiv.org/abs/2102.10336 · HTML: https://mlsanity.com/paper/2102.10336/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 12 references that are in the map.
- First posted on arXiv 20 Feb 2021.

## Builds on (most cited first, 11 of 11)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [SGDR: Stochastic Gradient Descent with Warm Restarts](https://mlsanity.com/paper/1608.03983/) (2016)
- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [Learning to Simulate Complex Physics with Graph Networks](https://mlsanity.com/paper/2002.09405/) (2020)
- [PHYRE: A New Benchmark for Physical Reasoning](https://mlsanity.com/paper/1908.05656/) (2019)
- [Learning Long-term Visual Dynamics with Region Proposal Interaction Networks](https://mlsanity.com/paper/2008.02265/) (2020)
- [Dynamics-aware Embeddings](https://mlsanity.com/paper/1908.09357/) (2019)
- [Forward Prediction for Physical Reasoning](https://mlsanity.com/paper/2006.10734/) (2020)

## Built on by (most cited first, 4 of 4)

- [SlotFormer: Unsupervised Visual Dynamics Simulation with Object-Centric Models](https://mlsanity.com/paper/2210.05861/) (2022)
- [Benchmarks for Physical Reasoning AI](https://mlsanity.com/paper/2312.10728/) (2023)
- [Causal-PIK: Causality-based Physical Reasoning with a Physics-Informed Kernel](https://mlsanity.com/paper/2505.22861/) (2025)
- [Bridging Physical Reasoning and Task Generalization via Visual Action Outcome Reasoning Alignment](https://mlsanity.com/paper/2607.06522/) (2026)

## Abstract (arXiv)

A common approach to solving physical reasoning tasks is to train a value learner on example tasks. A limitation of such an approach is that it requires learning about object dynamics solely from reward values assigned to the final state of a rollout of the environment. This study aims to address this limitation by augmenting the reward value with self-supervised signals about object dynamics. Specifically, we train the model to characterize the similarity of two environment rollouts, jointly with predicting the outcome of the reasoning task. This similarity can be defined as a distance measure between the trajectory of objects in the two rollouts, or learned directly from pixels using a contrastive formulation. Empirically, we find that this approach leads to substantial performance improvements on the PHYRE benchmark for physical reasoning (Bakhtin et al., 2019), establishing a new state-of-the-art.