aiqbench analysis
This paper introduces DLEPEM, a rehearsal-free framework for class-incremental learning that mitigates catastrophic forgetting by dynamically allocating task-specific LoRA experts and matching them with prototype ensembles. By combining frozen pre-trained prototypes with adaptive expert prototypes, the method effectively reduces cross-task interference while maintaining strong performance on standard benchmarks. The approach represents a significant advancement in parameter-efficient fine-tuning strategies for continuous machine learning scenarios.
Dynamic LoRA-Experts and Prototype-Ensemble Matching for Class-Incremental Learning
arXiv:2609.39839v1 Announce Type: new Abstract: Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-efficient fine-tuning with pre-trained models reduces parameter overhead but can suffer from cumulative interference and suboptimal alignment between inference samples and specialized modules. We propose Dynamic LoRA-Experts and Prototype-Ensemble Matching (DLEPEM), a two-stage rehearsal-free framework. DLEPEM allocates a task-specific LoRA-Expert for each incremental task to reduce cross-task interference, then combines frozen pre-trained-model prototypes with task-adaptive LoRA-Expert prototypes for reliable task-level discrimination. Experiments on standard CIL and Few-Shot CIL benchmarks demonstrate strong performance under the evaluated protocols.
Abstract:Class-Incremental Learning (CIL) aims to continuously learn new classes without forgetting previously acquired knowledge. Parameter-efficient fine-tuning with pre-trained models reduces parameter overhead but can suffer from cumulative interference and suboptimal alignment between inference samples and specialized modules. We propose Dynamic LoRA-Experts and Prototype-Ensemble Matching (DLEPEM), a two-stage rehearsal-free framework. DLEPEM allocates a task-specific LoRA-Expert for each incremental task to reduce cross-task interference, then combines frozen pre-trained-model prototypes with task-adaptive LoRA-Expert prototypes for reliable task-level discrimination. Experiments on standard CIL and Few-Shot CIL benchmarks demonstrate strong performance under the evaluated protocols.
Machine Learning (cs.LG)
arXiv:2609.39839 [cs.LG]
(or arXiv:2609.39839v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2609.39839 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.3390/app16126153 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 194 words
Related coverage
Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning
arXiv · Oct 1, 2026
Hyperbolic Prototype Routing for Rehearsal-Free Class-Incremental Learning
arXiv · Oct 1, 2026
Beyond Routing Saturation: A Long-Horizon Class-Incremental Perspective on Expert Routing in Multimodal Continual Instruction Tuning
arXiv · Aug 5, 2026
[AINews] Opus 5.5 is good at explainer videos
Latent Space · Sep 29, 2026