Plasticine: Accelerating Research in Plasticity-Motivated Deep Reinforcement Learning
Not on the map. Why?
Abstract (arXiv)
Developing lifelong learning agents is crucial for artificial general intelligence (AGI). However, deep reinforcement learning (RL) systems often suffer from plasticity loss, where neural networks gradually lose their ability to adapt during training. Despite its significance, this field lacks unified benchmarks and evaluation protocols. We introduce Plasticine, the first open-source framework for benchmarking plasticity optimization in deep RL. Plasticine provides single-file implementations of over 13 mitigation methods, 6 evaluation metrics, and learning scenarios with increasing non-stationarity levels from standard to continually varying environments. This framework enables researchers to systematically quantify plasticity loss, evaluate mitigation strategies, and analyze plasticity dynamics across different contexts. Our documentation, examples, and source code are available at https://github.com/RLE-Foundation/Plasticine.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 37 of its 49 references that are in the map. First posted on arXiv 24 Apr 2025.
Builds on
37 of 49 references.
References · 37
- 2017Proximal Policy Optimization Algorithmscited by 14,441
- 2016Layer Normalizationcited by 5,579
- 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,635
- 2018Addressing Function Approximation Error in Actor-Critic Methodscited by 2,302
- 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 1,916
- 2018DeepMind Control Suitecited by 935
- 2017A Distributional Perspective on Reinforcement Learningcited by 743
- 2019Leveraging Procedural Generation to Benchmark Reinforcement Learningcited by 478
- 2022The Primacy Bias in Deep Reinforcement Learningcited by 242
- 2023The Dormant Neuron Phenomenon in Deep Reinforcement Learningcited by 171
- 2023Understanding plasticity in neural networkscited by 170
- 2020Implicit Under-Parameterization Inhibits Data-Efficient Deep Reinforcement Learningcited by 134
25 more
- 2022Understanding and Preventing Capacity Loss in Reinforcement Learningcited by 124
- 2023Loss of Plasticity in Continual Deep Reinforcement Learningcited by 118
- 2022CORL: Research-oriented Deep Offline Reinforcement Learning Librarycited by 117
- 2024Bigger, Regularized, Optimistic: scaling for compute and sample-efficient continuous controlcited by 104
- 2023A Definition of Continual Reinforcement Learningcited by 98
- 2024Disentangling the Causes of Plasticity Loss in Neural Networkscited by 82
- 2024Craftax: A Lightning-Fast Benchmark for Open-Ended Reinforcement Learningcited by 80
- 2023Deep Reinforcement Learning with Plasticity Injectioncited by 71
- 2024Simplifying Deep Temporal Difference Learningcited by 71
- 2024SimBa: Simplicity Bias for Scaling Up Parameters in Deep Reinforcement Learningcited by 70
- 2021DR3: Value-Based Deep Reinforcement Learning Requires Explicit Regularizationcited by 66
- 2023Maintaining Plasticity in Continual Learning via Regenerative Regularizationcited by 63
- 2024Normalization and effective learning rates in reinforcement learningcited by 55
- 2023PLASTIC: Improving Input and Label Plasticity for Sample Efficient Reinforcement Learningcited by 42
- 2023Small batch deep reinforcement learningcited by 35
- 2025Hyperspherical Normalization for Scalable Deep Reinforcement Learningcited by 35
- 2023Revisiting Plasticity in Visual Reinforcement Learning: Data, Modules and Training Stagescited by 27
- 2024A Study of Plasticity Loss in On-Policy Deep Reinforcement Learningcited by 26
- 2024Plasticity Loss in Deep Reinforcement Learning: A Surveycited by 24
- 2024Parseval Regularization for Continual Reinforcement Learningcited by 21
- 2024Non-Stationary Learning of Neural Networks with Automatic Soft Parameter Resetcited by 17
- 2024Fast TRAC: A Parameter-Free Optimizer for Lifelong Reinforcement Learningcited by 16
- 2024Plastic Learning with Deep Fourier Featurescited by 14
- 2024Discounted Adaptive Online Learning: Towards Better Regularizationcited by 12
- 2025Activation by Interval-wise Dropout: A Simple Way to Prevent Neural Networks from Plasticity Losscited by 6
Built on by
Citing papers per year: 2025: 1; 2026: 1.
2 papers.