2026年5月6日提交的多篇研究论文探索了跨各种人工智能领域的持续学习的新方法。其中一篇论文介绍了一种基于重放的物理信息神经网络算子策略，以减轻灾难性遗忘。另一篇论文提出使用软令牌的“技能新词”来扩展大型语言模型的能力，而无需更新权重。此外，关于大型语言模型系统的研究提出了一种受生物记忆启发的、用于持续知识更新的多时间尺度记忆动力学方法。 AI
影响 这些论文探讨了提高人工智能在不忘记过去知识的情况下持续学习的能力的方法，这对于适应性和不断发展的系统至关重要。
排序理由 2026年5月6日发表的多篇arXiv论文，详细介绍了持续学习的新研究。
AI 生成摘要 · Google Gemini · 来自 19 个来源。 我们如何撰写摘要 →
arXiv:2605.05732v1 Announce Type: new Abstract: Large language models (LLMs) can acquire new capabilities through fine-tuning, but continual adaptation often leads to catastrophic forgetting. We propose CRAFT, a continual learning framework that avoids updating model weights by i…
arXiv:2605.05285v1 Announce Type: new Abstract: Large language models (LLMs) often suffer from catastrophic forgetting in continual learning: after learning new tasks sequentially, they perform worse on earlier tasks. Existing methods mitigate catastrophic forgetting by data repl…
arXiv:2605.04832v1 Announce Type: new Abstract: Neural operators generally demonstrate strong predictive performance on in-distribution (ID) problems. However, a critical limitation of existing methods is their significant performance degradation when encountering out-of-distribu…
arXiv:2309.09550v4 Announce Type: replace-cross Abstract: The human brain can self-organize rich and diverse sparse neural pathways to incrementally master hundreds of cognitive tasks. However, most existing continual learning algorithms for deep artificial and spiking neural net…
arXiv:2511.01553v2 Announce Type: replace Abstract: AI systems on edge devices require online continual learning -- adapting to non-stationary streams and unfamiliar classes without catastrophic forgetting -- under strict power constraints. We present CLP-SNN, a spiking neural ne…
arXiv:2605.05097v1 Announce Type: new Abstract: LLMs are trained once, then deployed into a world that never stops changing. External memory compensates for this, but most systems manage it explicitly rather than letting it adapt on its own. Biological memory works differently: c…
arXiv:2605.04970v1 Announce Type: new Abstract: Modern LLMs show mastery over an ever-growing range of skills, as well as the ability to compose them flexibly. However, extending model capabilities to new skills in a scalable manner is an open-problem: fine-tuning and parameter-e…
LLMs are trained once, then deployed into a world that never stops changing. External memory compensates for this, but most systems manage it explicitly rather than letting it adapt on its own. Biological memory works differently: coupled multi-timescale dynamics make new associa…
Modern LLMs show mastery over an ever-growing range of skills, as well as the ability to compose them flexibly. However, extending model capabilities to new skills in a scalable manner is an open-problem: fine-tuning and parameter-efficient variants risk catastrophic forgetting, …
Neural operators generally demonstrate strong predictive performance on in-distribution (ID) problems. However, a critical limitation of existing methods is their significant performance degradation when encountering out-of-distribution (OOD) data. To address this issue, this wor…
arXiv:2605.03866v1 Announce Type: new Abstract: Contrastive Language-Image Pretraining (CLIP) models excel at understanding image-text relationships but struggle with adapting to new data without forgetting prior knowledge. To address this, models are typically fine-tuned using b…
arXiv:2605.03085v1 Announce Type: new Abstract: Electroencephalography (EEG) signals provide millisecond-level temporal resolution but their analysis is limited by remarkable noise and inter-subject variability, making robust personalization difficult under limited annotations. U…
Contrastive Language-Image Pretraining (CLIP) models excel at understanding image-text relationships but struggle with adapting to new data without forgetting prior knowledge. To address this, models are typically fine-tuned using both new task data and a memory buffer of past ta…
arXiv:2605.01131v1 Announce Type: new Abstract: In continual reinforcement learning (CRL), good performance requires never-ending learning, acting, and exploration in a big, partially observable world. Most CRL experiments have focused on loss of plasticity -- the inability to ke…
arXiv:2605.02509v1 Announce Type: new Abstract: Continual learning systems face a fundamental tension between plasticity -- acquiring new knowledge -- and stability -- retaining prior knowledge. We introduce MPCS (Multi-Plasticity Continual System), a neuroplastic architecture th…
Continual learning systems face a fundamental tension between plasticity -- acquiring new knowledge -- and stability -- retaining prior knowledge. We introduce MPCS (Multi-Plasticity Continual System), a neuroplastic architecture that integrates eleven complementary mechanisms: t…
arXiv:2510.17281v5 Announce Type: replace Abstract: Scaling up data, parameters, and test-time computation has been the mainstream methods to improve LLM systems (LLMsys), but their upper bounds are almost reached due to the gradual depletion of high-quality data and marginal gai…
arXiv:2604.27003v1 Announce Type: cross Abstract: Memory-augmented LLM agents offer an appealing shortcut to continual learning: rather than updating model parameters, they accumulate experience in external memory, seemingly sidestepping the stability-plasticity dilemma of parame…
arXiv:2511.12090v2 Announce Type: replace Abstract: Prompt-based continual learning methods fine-tune only a small set of additional learnable parameters while keeping the pre-trained model's parameters frozen. It enables efficient adaptation to new tasks while mitigating the ris…