# Trajectory First: A Curriculum for Discovering Diverse Policies

arXiv: https://arxiv.org/abs/2506.01568 · HTML: https://mlsanity.com/paper/2506.01568/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 71 of its 107 references that are in the map.
- First posted on arXiv 2 Jun 2025.

## Builds on (most cited first, 30 of 71)

- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback](https://mlsanity.com/paper/2204.05862/) (2022)
- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [Addressing Function Approximation Error in Actor-Critic Methods](https://mlsanity.com/paper/1802.09477/) (2018)
- [Relational inductive biases, deep learning, and graph networks](https://mlsanity.com/paper/1806.01261/) (2018)
- [On the Continuity of Rotation Representations in Neural Networks](https://mlsanity.com/paper/1812.07035/) (2018)
- [Exploration by Random Network Distillation](https://mlsanity.com/paper/1810.12894/) (2018)
- [Diversity is All You Need: Learning Skills without a Reward Function](https://mlsanity.com/paper/1802.06070/) (2018)
- [Evolution Strategies as a Scalable Alternative to Reinforcement Learning](https://mlsanity.com/paper/1703.03864/) (2017)
- [Learning agile and dynamic motor skills for legged robots](https://mlsanity.com/paper/1901.08652/) (2019)
- [Illuminating search spaces by mapping elites](https://mlsanity.com/paper/1504.04909/) (2015)
- [Population Based Training of Neural Networks](https://mlsanity.com/paper/1711.09846/) (2017)
- [Efficient Online Reinforcement Learning with Offline Data](https://mlsanity.com/paper/2302.02948/) (2023)
- [Successor Features for Transfer in Reinforcement Learning](https://mlsanity.com/paper/1606.05312/) (2016)
- [Overcoming Exploration in Reinforcement Learning with Demonstrations](https://mlsanity.com/paper/1709.10089/) (2017)
- [Leveraging Demonstrations for Deep Reinforcement Learning on Robotics Problems with Sparse Rewards](https://mlsanity.com/paper/1707.08817/) (2017)
- [Dynamics-Aware Unsupervised Discovery of Skills](https://mlsanity.com/paper/1907.01657/) (2019)
- [Randomized Ensembled Double Q-Learning: Learning Fast Without a Model](https://mlsanity.com/paper/2101.05982/) (2021)
- [Noisy Networks for Exploration](https://mlsanity.com/paper/1706.10295/) (2017)
- [Cal-QL: Calibrated Offline RL Pre-Training for Efficient Online Fine-Tuning](https://mlsanity.com/paper/2303.05479/) (2023)
- [ManiSkill3: GPU Parallelized Robotics Simulation and Rendering for Generalizable Embodied AI](https://mlsanity.com/paper/2410.00425/) (2024)
- [Robots that can adapt like animals](https://mlsanity.com/paper/1407.3501/) (2014)
- [Behavior From the Void: Unsupervised Active Pre-Training](https://mlsanity.com/paper/2103.04551/) (2021)
- [Exploration in Deep Reinforcement Learning: A Survey](https://mlsanity.com/paper/2205.00824/) (2022)
- [Parameter Space Noise for Exploration](https://mlsanity.com/paper/1706.01905/) (2017)
- [Hybrid RL: Using Both Offline and Online Data Can Make RL Efficient](https://mlsanity.com/paper/2210.06718/) (2022)
- [Explore, Discover and Learn: Unsupervised Discovery of State-Covering Skills](https://mlsanity.com/paper/2002.03647/) (2020)
- [Effective Diversity in Population Based Reinforcement Learning](https://mlsanity.com/paper/2002.00632/) (2020)
- [Jump-Start Reinforcement Learning](https://mlsanity.com/paper/2204.02372/) (2022)
- [Improving Exploration in Evolution Strategies for Deep Reinforcement Learning via a Population of Novelty-Seeking Agents](https://mlsanity.com/paper/1712.06560/) (2017)

## Abstract (arXiv)

Being able to solve a task in diverse ways makes agents more robust to task variations and less prone to local optima. In this context, constrained diversity optimization has become a useful reinforcement learning (RL) framework for training a set of diverse agents in parallel. However, existing constrained-diversity RL methods often under-explore in complex tasks such as robot manipulation, resulting in limited behavioral diversity. We address this with a two-stage curriculum that introduces a spline-based trajectory prior as an inductive bias to produce diverse, high-reward behaviors in an initial stage, and then distills these behaviors into reactive, step-wise policies in a second stage. In our empirical evaluation, we provide novel insights into challenges of diversity-targeted training and show that our curriculum increases the diversity of learned skills while maintaining high task performance.