# Mirror Learning

arXiv: https://arxiv.org/abs/2607.28737 · HTML: https://mlsanity.com/paper/2607.28737/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 44 of its 58 references that are in the map.
- First posted on arXiv 30 Jul 2026.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Builds on (most cited first, 30 of 44)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets](https://mlsanity.com/paper/2311.15127/) (2023)
- [CARLA: An Open Urban Driving Simulator](https://mlsanity.com/paper/1711.03938/) (2017)
- [Video Diffusion Models](https://mlsanity.com/paper/2204.03458/) (2022)
- [Shortcut Learning in Deep Neural Networks](https://mlsanity.com/paper/2004.07780/) (2020)
- [RePaint: Inpainting using Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2201.09865/) (2022)
- [Learning Latent Dynamics for Planning from Pixels](https://mlsanity.com/paper/1811.04551/) (2018)
- [Curiosity-driven Exploration by Self-supervised Prediction](https://mlsanity.com/paper/1705.05363/) (2017)
- [End to End Learning for Self-Driving Cars](https://mlsanity.com/paper/1604.07316/) (2016)
- [WorldSimBench: Towards Video Generation Models as World Simulators](https://mlsanity.com/paper/2410.18072/) (2024)
- [End-to-End Training of Deep Visuomotor Policies](https://mlsanity.com/paper/1504.00702/) (2015)
- [Behavioral Cloning from Observation](https://mlsanity.com/paper/1805.01954/) (2018)
- [Time-Contrastive Networks: Self-Supervised Learning from Video](https://mlsanity.com/paper/1704.06888/) (2017)
- [UniPC: A Unified Predictor-Corrector Framework for Fast Sampling of Diffusion Models](https://mlsanity.com/paper/2302.04867/) (2023)
- [Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos](https://mlsanity.com/paper/2206.11795/) (2022)
- [Flexible Diffusion Modeling of Long Videos](https://mlsanity.com/paper/2205.11495/) (2022)
- [Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline](https://mlsanity.com/paper/2206.08129/) (2022)
- [World Simulation with Video Foundation Models for Physical AI](https://mlsanity.com/paper/2511.00062/) (2025)
- [COLMAP-Free 3D Gaussian Splatting](https://mlsanity.com/paper/2312.07504/) (2023)
- [Imitation from Observation: Learning to Imitate Behaviors from Raw Video via Context Translation](https://mlsanity.com/paper/1707.03374/) (2017)
- [Panoptic nuScenes: A Large-Scale Benchmark for LiDAR Panoptic Segmentation and Tracking](https://mlsanity.com/paper/2109.03805/) (2021)
- [Cosmos-Reason1: From Physical Common Sense To Embodied Reasoning](https://mlsanity.com/paper/2503.15558/) (2025)
- [Learning to Poke by Poking: Experiential Learning of Intuitive Physics](https://mlsanity.com/paper/1606.07419/) (2016)
- [SimLingo: Vision-Only Closed-Loop Autonomous Driving with Language-Action Alignment](https://mlsanity.com/paper/2503.09594/) (2025)
- [Alpamayo-R1: Bridging Reasoning and Action Prediction for Generalizable Autonomous Driving in the Long Tail](https://mlsanity.com/paper/2511.00088/) (2025)
- [Learning from All Vehicles](https://mlsanity.com/paper/2203.11934/) (2022)
- [Third-Person Imitation Learning](https://mlsanity.com/paper/1703.01703/) (2017)
- [DiffuEraser: A Diffusion Model for Video Inpainting](https://mlsanity.com/paper/2501.10018/) (2025)
- [WOD-E2E: Waymo Open Dataset for End-to-End Driving in Challenging Long-tail Scenarios](https://mlsanity.com/paper/2510.26125/) (2025)

## Abstract (arXiv)

We investigate imitation learning through the lens of third-person observation and propose a framework for mirror learning: acquiring actionable policies from passive observation. While behavior cloning (BC) excels under dense, well-aligned first-person data, it fundamentally fails to leverage the rich observational signals arising from third-person demonstrations that humans and animals routinely exploit. We introduce a method that composes (i) a learned perspective transformation that places learners in demonstrators' shoes using a fine-tuned video diffusion model and (ii) an inverse dynamics model that infers action trajectories in the learners' control space. This enables the synthesis of mirror data, pseudo first-person expert data generated from third-person observations of demonstrator behavior. Empirically, we show that mirror data alone can train effective policies, and that augmenting first-person BC training with mirror data further improves downstream policy performance. Our results suggest that modern generative world models implicitly encode sufficient structure to enable a scalable and safe alternative to teleoperation-heavy data collection.