| title | Train Smarter, Not Harder: Switching Signal-Guided Training in Active Learning |
|---|---|
| authors | Nagham Omar, Maya Rozenshtein, Evgeny Mishlyakov, Avigdor Gal |
| date | 2026-09-06 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2609.06806 |
| affiliation | Technion Israel institute of technology |
| upvotes | 5 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Training strategy, namely whether to retrain from scratch or fine-tune from the previous checkpoint, is an overlooked decision variable in active learning. We show that this choice has exploitable structure: retraining is most useful in early rounds, when each batch can substantially reshape the labeled distribution, while fine-tuning becomes safer once the model trajectory stabilizes. We propose HybridAL, an adaptive training schedule that monitors an online stabilization signal and switches from retraining to fine-tuning after sustained stabilization. Two complementary signals, spectral exponent change Δα (weight-based) and accuracy change ΔAcc (validation-based), span different points on the time-calibration trade-off. Across three encoder backbones and six text-classification tasks (five seeds each), HybridAL keeps endpoint macro-F1 non-inferior to retraining and fine-tuning at a 0.010 margin, saves up to 49% of retraining time, and recovers a substantial fraction of retraining's calibration advantage as measured by negative log-likelihood (NLL). Compared with schedules that switch at a pre-committed round, HybridAL obtains lower NLL at moderate additional cost, showing that trajectory-dependent switching provides a stronger time-calibration trade-off than fixed early switching. |
cite
@misc{omar2026train,
title={Train Smarter, Not Harder: Switching Signal-Guided Training in Active Learning},
author={Nagham Omar and Maya Rozenshtein and Evgeny Mishlyakov and Avigdor Gal},
year={2026},
eprint={2609.06806},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2609.06806}
}