# TAR: Teacher-Aligned Representations via Contrastive Learning for Quadrupedal Locomotion

arXiv: https://arxiv.org/abs/2503.20839 · HTML: https://mlsanity.com/paper/2503.20839/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in On-policy and self-distillation for LLMs, one of 12 directions in Knowledge distillation.
- Builds on 28 of its 36 references that are in the map.
- First posted on arXiv 26 Mar 2025.

## Placed in

- Knowledge distillation / [On-policy and self-distillation for LLMs](https://mlsanity.com/distill/on-policy-and-self-distillation-for-llms/)
- Deep reinforcement learning methods / [RL for robot locomotion and manipulation](https://mlsanity.com/deeprl/rl-for-robot-locomotion-and-manipulation/)

## Builds on (most cited first, 28 of 28)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [A Simple Framework for Contrastive Learning of Visual Representations](https://mlsanity.com/paper/2002.05709/) (2020)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Unsupervised Learning of Visual Features by Contrasting Cluster Assignments](https://mlsanity.com/paper/2006.09882/) (2020)
- [CURL: Contrastive Unsupervised Representations for Reinforcement Learning](https://mlsanity.com/paper/2004.04136/) (2020)
- [Learning Quadrupedal Locomotion over Challenging Terrain](https://mlsanity.com/paper/2010.11251/) (2020)
- [RMA: Rapid Motor Adaptation for Legged Robots](https://mlsanity.com/paper/2107.04034/) (2021)
- [Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillation](https://mlsanity.com/paper/1905.08094/) (2019)
- [Learning robust perceptive locomotion for quadrupedal robots in the wild](https://mlsanity.com/paper/2201.08117/) (2022)
- [Challenges of Real-World Reinforcement Learning](https://mlsanity.com/paper/1904.12901/) (2019)
- [Reinforcement Learning with Prototypical Representations](https://mlsanity.com/paper/2102.11271/) (2021)
- [Rapid Locomotion via Reinforcement Learning](https://mlsanity.com/paper/2205.02824/) (2022)
- [Walk These Ways: Tuning Robot Control for Generalization with Multiplicity of Behavior](https://mlsanity.com/paper/2212.03238/) (2022)
- [Reinforcement Learning for Versatile, Dynamic, and Robust Bipedal Locomotion Control](https://mlsanity.com/paper/2401.16889/) (2024)
- [DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representations](https://mlsanity.com/paper/2110.14565/) (2021)
- [DreamWaQ: Learning Robust Quadrupedal Locomotion With Implicit Terrain Imagination via Deep Reinforcement Learning](https://mlsanity.com/paper/2301.10602/) (2023)
- [Data Efficient Reinforcement Learning for Legged Robots](https://mlsanity.com/paper/1907.03613/) (2019)
- [Not Only Rewards But Also Constraints: Applications on Legged Robot Locomotion](https://mlsanity.com/paper/2308.12517/) (2023)
- [Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response](https://mlsanity.com/paper/2312.11460/) (2023)
- [Learning Memory-Based Control for Human-Scale Bipedal Locomotion](https://mlsanity.com/paper/2006.02402/) (2020)
- [How to pick the domain randomization parameters for sim-to-real transfer of reinforcement learning policies?](https://mlsanity.com/paper/1903.11774/) (2019)
- [CTS: Concurrent Teacher-Student Reinforcement Learning for Legged Locomotion](https://mlsanity.com/paper/2405.10830/) (2024)
- [DROPO: Sim-to-Real Transfer with Offline Domain Randomization](https://mlsanity.com/paper/2201.08434/) (2022)
- [On Covariate Shift of Latent Confounders in Imitation and Reinforcement Learning](https://mlsanity.com/paper/2110.06539/) (2021)
- [PA-LOCO: Learning Perturbation-Adaptive Locomotion for Quadruped Robots](https://mlsanity.com/paper/2407.04224/) (2024)
- [SLR: Learning Quadruped Locomotion without Privileged Information](https://mlsanity.com/paper/2406.04835/) (2024)
- [Learning Bipedal Walking on a Quadruped Robot via Adversarial Motion Priors](https://mlsanity.com/paper/2407.02282/) (2024)

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

- [To Distill or Decide? Understanding the Algorithmic Trade-off in Partially Observable Reinforcement Learning](https://mlsanity.com/paper/2510.03207/) (2025)
- [TRANS: Terrain-aware Reinforcement Learning for Agile Navigation of Quadruped Robots under Social Interactions](https://mlsanity.com/paper/2602.12724/) (2026)
- [DAWN: Noise-Robust Quadruped Parkour via Depth-Denoising World Models](https://mlsanity.com/paper/2609.29092/) (2026)
- [PROMO: Preference-conditioned Multi-Objective Reinforcement Learning for Quadrupedal Robots](https://mlsanity.com/paper/2610.01260/) (2026)
- [Transporting Unsecured Stacked Payloads with a Quadrupedal Robot via Multi-Objective Reinforcement Learning](https://mlsanity.com/paper/2610.05819/) (2026)

## Abstract (arXiv)

Quadrupedal locomotion via Reinforcement Learning (RL) is commonly addressed using the teacher-student paradigm, where a privileged teacher guides a proprioceptive student policy. However, key challenges such as representation misalignment between privileged teacher and proprioceptive-only student, covariate shift due to behavioral cloning, and lack of deployable adaptation; lead to poor generalization in real-world scenarios. We propose Teacher-Aligned Representations via Contrastive Learning (TAR), a framework that leverages privileged information with self-supervised contrastive learning to bridge this gap. By aligning representations to a privileged teacher in simulation via contrastive objectives, our student policy learns structured latent spaces and exhibits robust generalization to Out-of-Distribution (OOD) scenarios, surpassing the fully privileged "Teacher". Results showed accelerated training by 2x compared to state-of-the-art baselines to achieve peak performance. OOD scenarios showed better generalization by 40% on average compared to existing methods. Moreover, TAR transitions seamlessly into learning during deployment without requiring privileged states, setting a new benchmark in sample-efficient, adaptive locomotion and enabling continual fine-tuning in real-world scenarios. Open-source code and videos are available at https://amrmousa.com/TARLoco/.