mlsanityData to 7 Oct 2026

GR4CIL: Gap-compensated Routing for CLIP-based Class Incremental Learning

arXiv 2604.17822 ↗First posted 20 Apr 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Class-Incremental Learning (CIL) aims to continuously acquire new categories while preserving previously learned knowledge. Recently, Contrastive Language-Image Pre-trained (CLIP) models have shown strong potential for CIL due to their powerful generalization ability. However, existing methods still face two key challenges: shared-parameter adaptation tends to cause old-knowledge drift, and task-specific knowledge organization often leads to poorly calibrated cross-task responses, making reliable routing difficult. To address these issues, we propose GR4CIL, a framework combining task discrimination and knowledge routing for CLIP-based CIL. GR4CIL preserves task-specific visual knowledge while maintaining an incrementally stable shared textual semantic space, thereby reducing interference across tasks. Moreover, we introduce an orthogonal compensation mechanism to mitigate modality-gap-induced bias, enhance within-task discrimination, and enlarge the score margin between the ground-truth task and competing tasks. As a result, GR4CIL enables more reliable task-aware routing over learned knowledge while retaining the zero-shot generalization capability. Experiments on multiple benchmarks show that GR4CIL consistently outperforms strong baselines.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 38 of its 48 references that are in the map. First posted on arXiv 20 Apr 2026.

Builds on

38 of 48 references.

References · 38

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  3. 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
  4. 2021Learning to Prompt for Vision-Language Modelscited by 2,269
  5. 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,842
  6. 2021CLIP-Adapter: Better Vision-Language Models with Feature Adapterscited by 1,021
  7. 2021Learning to Prompt for Continual Learningcited by 938
  8. 2022DualPrompt: Complementary Prompting for Rehearsal-free Continual Learningcited by 635
  9. 2022Mind the Gap: Understanding the Modality Gap in Multi-modal Contrastive Representation Learningcited by 553
  10. 2022CODA-Prompt: COntinual Decomposed Attention-based Prompting for Rehearsal-Free Continual Learningcited by 446
  11. 2024Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapterscited by 199
  12. 2023Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Needcited by 183
26 more
  1. 2023SLCA: Slow Learner with Classifier Alignment for Continual Learning on a Pre-trained Modelcited by 178
  2. 2022A Model or 603 Exemplars: Towards Memory-Efficient Class-Incremental Learningcited by 152
  3. 2022How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?cited by 128
  4. 2023Preventing Zero-Shot Transfer Degradation in Continual Learning of Vision-Language Modelscited by 112
  5. 2023Learning without Forgetting for Vision-Language Modelscited by 79
  6. 2021Provable Guarantees for Understanding Out-of-distribution Detectioncited by 78
  7. 2022A Theoretical Study on Solving Continual Learningcited by 78
  8. 2024MagMax: Leveraging Model Merging for Seamless Continual Learningcited by 67
  9. 2022CLIP model is an Efficient Continual Learnercited by 64
  10. 2025SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learningcited by 60
  11. 2024Two Effects, One Trigger: On the Modality Gap, Object Bias, and Information Imbalance in Contrastive Vision-Language Modelscited by 55
  12. 2024Class-Incremental Learning with CLIP: Adaptive Representation Adjustment and Parameter Fusioncited by 47
  13. 2024Learning with Mixture of Prototypes for Out-of-Distribution Detectioncited by 43
  14. 2025Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversioncited by 42
  15. 2022A Multi-Head Model for Continual Learning via Out-of-Distribution Replaycited by 27
  16. 2024CLAP4CLIP: Continual Learning with Probabilistic Finetuning for Vision-Language Modelscited by 25
  17. 2022Continual Learning Based on OOD Detection and Task Maskingcited by 21
  18. 2025Post-pre-training for Modality Alignment in Vision-Language Foundation Modelscited by 18
  19. 2023Class Incremental Learning via Likelihood Ratio Based Task Predictioncited by 17
  20. 2025LADA: Scalable Label-Specific CLIP Adapter for Continual Learningcited by 11
  21. 2025External Knowledge Injection for CLIP-Based Class-Incremental Learningcited by 10
  22. 2024Towards Continual Learning Desiderata via HSIC-Bottleneck Orthogonalization and Equiangular Embeddingcited by 9
  23. 2025On the Value of Cross-Modal Misalignment in Multimodal Representation Learningcited by 9
  24. 2025On the Discrimination and Consistency for Exemplar-Free Class Incremental Learningcited by 8
  25. 2024Harnessing Neural Unit Dynamics for Effective and Scalable Class-Incremental Learningcited by 4
  26. 2025Semantic-guided LoRA Parameters Generationcited by 1

Built on by

Not cited yet.