ICML 2026 accepted papers on Continual learning
98 of the 6,341 papers accepted at ICML 2026 (1.5%) are on Continual learning, compared with 1.5% at ICML 2025. They fall into 4 topics.
See ICML 2026 in the interactive map
Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.
Topics and papers
Continual fine-tuning of LLMs · 53 papers
- Can Muon Fine-tune Adam-Pretrained Models?
- Continual Model Routing in Evolving Model Hubs
- DualOptim+: Bridging Shared and Decoupled Optimizer States for Better Machine Unlearning in Large Language Models
- ECA: Efficient Continual Alignment for Open-Ended Image-to-Text Generation
- Energy-Structured Low-Rank Adaptation for Continual Learning
- Forget by Uncertainty: Orthogonal Entropy Unlearning for Quantized Neural Networks
- Forget to Know, Remember to Use: Context-Aware Unlearning for Large Language Models
- Hyperbolic Multimodal Continual Learning
- JANUS-LORA: A Balanced Low-Rank Adaptation for Continual Learning
- Just-In-Time Reinforcement Learning: Continual Learning in LLM Agents Without Gradient Updates
- KORE: Enhancing Knowledge Injection for Large Multimodal Models via Knowledge-Oriented Controls
- Little By Little: Continual Learning via Incremental Mixture of Rank-1 Associative Memory Experts
and 41 more in the interactive map.
Online and replay-based · 28 papers
- Active Continual Learning with Metaplastic Binary Bayesian Neural Networks
- Balancing Plasticity and Stability with Fast and Slow Successor Features
- CE$^4$L: Continual Ego, Exo, and Ego-Exo Learning
- Continual Learning through Control Minimization
- Learning to Remember, Learn, and Forget in Attention-Based Models
- MedCRP-CL: Continual Medical Image Segmentation via Bayesian Nonparametric Semantic Modality Discovery
- MePo: Meta Post-Refinement for Rehearsal-Free General Continual Learning
- Online Continual Learning with Dynamic Label Hierarchies
- Preserving Plasticity in Continual Learning via Dynamical Isometry
- RAG without Forgetting: Continual Query-Infused Key Memory
- Shapley Neuron Values for Continual Learning: Which Neurons Matter Most?
- SPHERE: Mitigating the Loss of Spectral Plasticity in Mixture-of-Experts for Deep Reinforcement Learning
and 16 more in the interactive map.
Class-incremental · 11 papers
- AREA: Attribute Extraction and Aggregation for CLIP-Based Class-Incremental Learning
- Beyond Point-wise Neural Collapse: A Topology-Aware Hierarchical Classifier for Class-Incremental Learning
- Breaking the Synthetic-Real Domain Shortcut for Training-Free Generative Replay-based Class Incremental Learning
- Focus, Align, and Sustain: Counteracting Gradient Dilution in Incremental Object Detection
- Symbiosis-Inspired Knowledge Distillation for Incremental Object Detection
- Cross-View Lewis Weight Fusion Empowering Exemplar Replay for Federated Class-Incremental Learning
- Forgetting Whenever You Want: A Decentralized Continual Learning Framework with On-Demand Unlearning
- Learnability-Driven Knowledge Assimilation for Class-Incremental Semantic Segmentation
- On the Power of Statistics in Class-Incremental Learning with Pretrained Models
- Persistent Backdoor Attacks in Class-Incremental Learning via Structural Invariant Anchoring
- Subspace-Aware Feature Reshaping for Open-Set Graph Class-Incremental Learning
Domain and prompt-based · 6 papers
- Continual Learning of Domain-Invariant Representations
- Continual Segmentation under Joint Nonstationarity
- HEDP: A Hybrid Energy-Distance Prompt-based Framework for Domain Incremental Learning
- OptProver: Bridging Olympiad and Optimization through Continual Training in Formal Theorem Proving
- Adversarial Latent Embedding Repair for LLM Continual Learning
- Parameter-Masked Decoupled Optimization for Cross-Domain Class-Incremental Learning