CVPR 2025 accepted papers on Continual learning

50 of the 2,871 papers accepted at CVPR 2025 (1.7%) are on Continual learning, compared with 1.8% at CVPR 2024. They fall into 4 topics.

See CVPR 2025 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

Class-incremental · 17 papers

Task-Agnostic Guided Feature Expansion for Class-Incremental Learning (17 citations)
Reducing Class-wise Confusion for Incremental Learning with Disentangled Manifolds (7 citations)
Revisiting Generative Replay for Class Incremental Object Detection (7 citations)
Dynamic Integration of Task-Specific Adapters for Class Incremental Learning (6 citations)
Knowledge Memorization and Rumination for Pre-trained Model-based Class-Incremental Learning (6 citations)
Adapter Merging with Centroid Prototype Mapping for Scalable Class-Incremental Learning (5 citations)
SEC-Prompt:SEmantic Complementary Prompting for Few-Shot Class-Incremental Learning (5 citations)
Beyond Background Shift: Rethinking Instance Replay in Continual Semantic Segmentation (4 citations)
FALCON: Fairness Learning via Contrastive Attention Approach to Continual Semantic Scene Understanding (4 citations)
Rethinking Query-based Transformer for Continual Image Segmentation (4 citations)
Activating Sparse Part Concepts for 3D Class Incremental Learning (3 citations)
CoMBO: Conflict Mitigation via Branched Optimization for Class Incremental Segmentation (3 citations)

and 5 more in the interactive map.

Continual fine-tuning of LLMs · 13 papers

CL-LoRA: Continual Low-Rank Adaptation for Rehearsal-Free Class-Incremental Learning (23 citations)
CL-MoE: Enhancing Multimodal Large Language Model with Dual Momentum Mixture-of-Experts for Continual Visual Question Answering (18 citations)
Self-Expansion of Pre-trained Models with Mixture of Adapters for Continual Learning (18 citations)
LoRA Subtraction for Drift-Resistant Space in Exemplar-Free Continual Learning (12 citations)
BiLoRA: Almost-Orthogonal Parameter Spaces for Continual Learning (10 citations)
Language Guided Concept Bottleneck Models for Interpretable Continual Learning (10 citations)
Decoupled Distillation to Erase: A General Unlearning Method for Any Class-centric Tasks (7 citations)
LoRASculpt: Sculpting LoRA for Harmonizing General and Specialized Knowledge in Multimodal Large Language Models (6 citations)
Synthetic Data is an Elegant GIFT for Continual Vision-Language Models (5 citations)
AdaDARE-gamma: Balancing Stability and Plasticity in Multi-modal LLMs through Efficient Adaptation (2 citations)
ESC: Erasing Space Concept for Knowledge Deletion (1 citation)
Low-Rank Adaptation in Multilinear Operator Networks for Security-Preserving Incremental Learning (1 citation)

and 1 more in the interactive map.

Online and replay-based · 13 papers

One-for-More: Continual Diffusion Model for Anomaly Detection (24 citations)
Continuous Adverse Weather Removal via Degradation-Aware Distillation (13 citations)
Stacking Brick by Brick: Aligned Feature Isolation for Incremental Face Forgery Detection (10 citations)
Handling Spatial-Temporal Data Heterogeneity for Federated Continual Learning via Tail Anchor (6 citations)
Advancing Multiple Instance Learning with Continual Learning for Whole Slide Imaging (5 citations)
KAC: Kolmogorov-Arnold Classifier for Continual Learning (5 citations)
ConceptGuard: Continual Personalized Text-to-Image Generation with Forgetting and Confusion Mitigation (3 citations)
H2ST: Hierarchical Two-Sample Tests for Continual Out-of-Distribution Detection (2 citations)
Do Your Best and Get Enough Rest for Continual Learning (1 citation)
Enhancing Online Continual Learning with Plug-and-Play State Space Model and Class-Conditional Mixture of Discretization (1 citation)
Online Task-Free Continual Learning via Dynamic Expansionable Memory Distribution (1 citation)
Towards Continual Universal Segmentation (1 citation)