Library · 11-persistent-memory-continual-adaptation
Persistent memory
| Title | Peer | Link |
|---|---|---|
| It's All Connected (Miras) | ◦ preprint | arxiv.org/abs/2504.13173 |
| Atlas: Optimally Memorize the Context at Test Time | ◦ preprint | arxiv.org/abs/2505.23735 |
| Gated Delta Networks (Mamba2 + delta rule) | ◦ preprint | arxiv.org/abs/2412.06464 |
| Test-Time Training with Next-Token Prediction (TTT-NTP) | ◦ preprint | arxiv.org/abs/2606.21803 |
| Lattice: Learning to Efficiently Compress the Memory | ◦ preprint | arxiv.org/abs/2504.05646 |
| SR-TTT: Surprisal-Aware Residual Test-Time Training | ◦ preprint | arxiv.org/abs/2603.06642 |
| Selective Memory: Write-Time Gating w/ Hierarchical Archiving | ◦ preprint | arxiv.org/abs/2603.15994 |
| Loss of Plasticity in Deep Continual Learning | ✓ peer | nature.com/articles/s41586-024-07711-7 |
| loss-of-plasticity (official repo, CBP + benchmarks) | github.com/shibhansh/loss-of-plasticity | |
| AdaLin: Preserving Plasticity via Adaptive Linearity Injection | ◦ preprint | arxiv.org/abs/2505.09486 |
| L2 Init: Maintaining Plasticity via Regenerative Regularization | ◦ preprint | arxiv.org/abs/2308.11958 |
| Plasticity Loss in Deep RL: A Survey | ◦ preprint | arxiv.org/abs/2411.04832 |
| Plasticine: Plasticity-Motivated Deep RL benchmark | ◦ preprint | arxiv.org/abs/2504.17490 |
| AltNet: Plasticity-Stability Dilemma in RL | — unref | arxiv.org/abs/2512.01034 |
| Mem0: Scalable Long-Term Memory for AI Agents | ◦ preprint | arxiv.org/abs/2504.19413 |
| A-MEM: Agentic Memory for LLM Agents | ◦ preprint | arxiv.org/abs/2502.12110 |
| LoCoMo: Evaluating Very-Long-Term Conversational Memory | aclanthology.org/2024.acl-long.747/ | |
| HebbGate: Local Reward-Modulated Gating for Continual Learning | — | |
| Computational principles of synaptic memory consolidation (Benna & Fusi) | ✓ peer | doi:10.1038/nn.4401 |
| Enabling Continual Learning with Differentiable Hebbian Plasticity (Thangarasa, Miconi) | — unref | arxiv.org/abs/2006.16558 |
| AGMP — Astrocyte-Gated Multi-Timescale Plasticity | ✓ peer | doi.org/10.3389/fnins.2025.1768235 |
| Hippocampus supports multi-task RL under partial observability | ✓ peer | nature.com/articles/s41467-025-64591-9 |
| MESU — Bayesian continual learning & forgetting (Bonnet et al.) | ◦ preprint | arxiv.org/abs/2504.13569 |
| Hybrid NNs for continual learning inspired by corticohippocampal circuits | — | |
| Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention (Katharopoulos et al., ICML 2020) | ✓ peer | arxiv.org/abs/2006.16236 |
| Linear Transformers Are Secretly Fast Weight Programmers (Schlag, Irie, Schmidhuber, ICML 2021) | ✓ peer | arxiv.org/abs/2102.11174 |
| Learning to Control Fast-Weight Memories (Schmidhuber) | ✓ peer | doi:10.1162/neco.1992.4.1.131 |
| Parallelizing Linear Transformers with the Delta Rule over Sequence Length — DeltaNet (Yang et al., NeurIPS 2024) | ◦ preprint | arxiv.org/abs/2406.06484 |
| Mean teachers are better role models (Tarvainen & Valpola, NeurIPS 2017) | ✓ peer | arxiv.org/abs/1703.01780 |
| Memory-R1: Enhancing LLM Agents to Manage & Utilize Memories via RL (Yan et al.) | ◦ preprint | arxiv.org/abs/2508.19828 |
| DeltaMem: Towards Agentic Memory Management via RL (Zhang et al.) | ◦ preprint | arxiv.org/abs/2604.01560 |
| Just-In-Time RL: Continual Learning in LLM Agents Without Gradient Updates (Li et al.) | ◦ preprint | arxiv.org/abs/2601.18510 |
| Reinforced Fast Weights with Next-Sequence Prediction (REFINE) | ◦ preprint | arxiv.org/abs/2602.16704 |
| Reliability-Adjusted Prioritized Experience Replay (ReaPER) | ◦ preprint | arxiv.org/abs/2506.18482 |
| Governing Evolving Memory in LLM Agents (SSGM Framework) | ◦ preprint | arxiv.org/abs/2603.11768 |
| AURA: Action-Gated Memory for Robot Policies at Constant VRAM (Chen et al.) | ◦ preprint | arxiv.org/abs/2606.02775 |
| Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention (Hatamizadeh et al.) | ◦ preprint | arxiv.org/abs/2605.22791 |
| Worth Remembering: Surprise-Gated Robot Episodic Memory (Gorlo et al.) | ◦ preprint | arxiv.org/abs/2606.03787 |
| Context-Gated Associative Retrieval: From Theory to Transformers (Choraria et al.) | ◦ preprint | arxiv.org/abs/2605.10970 |
| Short-Term Synaptic Plasticity Stabilizes Goal-Conditioned Dynamics (PFC reservoir) | ◦ preprint | arxiv.org/abs/2606.03481 |
| Continual Learning Through Synaptic Intelligence | ◦ preprint | arxiv.org/abs/1703.04200 |
| Gradient Episodic Memory for Continual Learning (GEM) | ✓ peer | arxiv.org/abs/1706.08840 |
| Efficient Lifelong Learning with A-GEM | ◦ preprint | arxiv.org/abs/1812.00420 |
| Prioritized Experience Replay (PER) | ◦ preprint | arxiv.org/abs/1511.05952 |
| Continual lifelong learning with neural networks: a review | ✓ peer | doi.org/10.1016/j.neunet.2019.01.012 |
| Overcoming catastrophic forgetting in neural networks (EWC) | ✓ peer | doi.org/10.1073/pnas.1611835114 |
| Catastrophic Interference in Connectionist Networks | ✓ peer | doi.org/10.1016/S0079-7421(08)60536-8 |
| Beyond Perplexity: A Behavioral Evaluation Framework for Deployment-Memory Claims in LLM Test-Time Training | ◦ preprint | arxiv.org/abs/2607.00368 |
| Can Scale Save Us From Plasticity Loss in Large Language Models? | ◦ preprint | arxiv.org/abs/2606.24752 |
| Fast and Slow Variational Continual Learning | ◦ preprint | arxiv.org/abs/2606.24007 |
| Can a Language Model Learn Facts Continually in Its Weights? | ◦ preprint | arxiv.org/abs/2607.11020 |
| Learning on the Job: Continual Learning from Deployment Feedback for Frozen-Weights Agents | ◦ preprint | arxiv.org/abs/2607.22157 |
| No Time Like the Present: Agentic Test-Time Training for LLM Agents | ◦ preprint | arxiv.org/abs/2607.03441 |
| Is Our Benchmark Enough? An Analysis of Continual Learning for MLLMs | ◦ preprint | arxiv.org/abs/2606.20961 |
| Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments | ◦ preprint | arxiv.org/abs/2606.05661 |
| LatentGym: A Testbed For Cross-Task Experiential Learning With Controllable Latent Structure | ◦ preprint | arxiv.org/abs/2606.15306 |
| Rethinking Continual Experience Internalization for Self-Evolving LLM Agents | ◦ preprint | arxiv.org/abs/2606.04703 |
| A Decision-Theoretic View of Test-Time Training: When, How Far, and Which Directions to Adapt | ◦ preprint | arxiv.org/abs/2606.15569 |
| Scaling Self-Evolving Agents via Parametric Memory (TMEM) | ◦ preprint | arxiv.org/abs/2606.04536 |
| Retrievable Gradients: Continual Post-Training Without Cumulative Weight Drift | ◦ preprint | arxiv.org/abs/2606.15734 |
| The Weight of Silence: A Causal Case for Weights Over the Scratchpad in Latent Chess Reasoning | arxiv.org/abs/2607.20952 |