Metraipolis › archive › dc_db665d80aeb0

Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning — arXiv:2609.39390

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.39390
License (as archived)
CC0-1.0
Archived
2026-10-07T13:46:09.449Z
SHA-256 of the archived text
d9d3cf02d4fd629fa8f57168cee96aa4a66d9dcbdfbccb0f9c039498e7a41459
Stable URL of this page
https://metraipolis.com/documento/dc_db665d80aeb0

How to cite

Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning — arXiv:2609.39390 — arXiv (en), https://arxiv.org/abs/2609.39390 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_db665d80aeb0 (Metraipolis archive dc_db665d80aeb0, sha256 d9d3cf02d4fd629fa8f57168cee96aa4a66d9dcbdfbccb0f9c039498e7a41459)
BibTeX
@misc{dcdb665d80aeb0,
  title = {Decoupled and Distilled: Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer for Few-Shot Class-Incremental Learning — arXiv:2609.39390},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.39390},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_db665d80aeb0; sha256 d9d3cf02d4fd629fa8f57168cee96aa4a66d9dcbdfbccb0f9c039498e7a41459},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39390. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Few-Shot Class-Incremental Learning (FSCIL) addresses the challenge of learning new classes from very limited samples while retaining knowledge of previously learned ones. Although parameter-efficient fine-tuning methods with pre-trained models show promise for class-incremental learning, strict gradient-based constraints can be unreliable under severe data scarcity, while multi-expert approaches can impose substantial inference-time costs. We propose TALON (Task-Adaptive LoRA-Teachers with Ensemble Knowledge Transfer), an inference-efficient FSCIL framework. TALON dynamically allocates an independent LoRA-Teacher to each incremental task for task-specific representation learning, then distills multiple frozen teachers into a unified LoRA-Student through Ensemble Knowledge Transfer, eliminating runtime module selection or generation. A semantic-guided distillation strategy weights teacher contributions by feature-space similarity to mitigate catastrophic forgetting and overfitting. Across three class-order runs, TALON achieves comparable or better mean average accuracy across four FSCIL benchmarks, obtaining 86.68 +/- 1.22% on CUB200, 90.39 +/- 0.27% on CIFAR100, 78.38 +/- 0.94% on ImageNet-R, and 96.34 +/- 0.33% on miniImageNet. TALON uses up to 33x fewer deployment parameters and reduces average inference time per task to 26.7 s, a 41.70% reduction relative to ASP.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.39390 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T13:18:10.376Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).