analiza aiqbench
This arXiv paper treats active learning and pseudo-labeling as a training-budget problem: under a fixed number of gradient steps per round, how should unlabeled, pseudo-labeled examples share mini-batches with labeled data? The proposed SSCAL method admits pseudo-labels based on distance to labeled neighbors, caps their share of each batch, and reweights the labeled loss to preserve its expected gradient contribution. Readers should note that the evidence is limited to two audio datasets and two frozen encoders, so the claimed ranking may not transfer to other modalities or to fine-tuned models.
Should Pseudo-Labels Share Mini-Batches with Labeled Examples? Semi-Supervised Continual Active Learning with Limited Training Cost
arXiv:2610.35345v1 Announce Type: new Abstract: Active learning (AL) reduces annotation by querying informative examples. Continual active learning (CAL) also cuts the training cost, the total number of gradient steps, by training one model across AL rounds with a fixed number of mini-batch steps per round. Pseudo-labels add supervision without annotation, but they increase training cost. In CAL, where every round trains on a fixed number of mini-batches, whether and how pseudo-labeled examples should share these mini-batches with labeled examples is therefore open. We provide the first formalization of this problem as two decisions after each query, namely (i) which unlabeled examples should enter training with a pseudo-label and (ii) what share of each mini-batch they should take. Our method, semi-supervised continual active learning (SSCAL), works with any acquisition function. It admits pseudo-labels based on distances from unlabeled examples to labeled neighbors, caps their share, and reweights the loss on labeled examples to keep their expected gradient contribution unchanged. On frozen embeddings, its pseudo-labeler updates incrementally without extra model forward passes. Across two audio datasets, two frozen encoders, six acquisition functions and ten seeds, SSCAL ranks first on average among four protocols with equal training steps, answering the question affirmatively.
Title:Should Pseudo-Labels Share Mini-Batches with Labeled Examples? Semi-Supervised Continual Active Learning with Limited Training Cost
Abstract:Active learning (AL) reduces annotation by querying informative examples. Continual active learning (CAL) also cuts the training cost, the total number of gradient steps, by training one model across AL rounds with a fixed number of mini-batch steps per round. Pseudo-labels add supervision without annotation, but they increase training cost. In CAL, where every round trains on a fixed number of mini-batches, whether and how pseudo-labeled examples should share these mini-batches with labeled examples is therefore open. We provide the first formalization of this problem as two decisions after each query, namely (i) which unlabeled examples should enter training with a pseudo-label and (ii) what share of each mini-batch they should take. Our method, semi-supervised continual active learning (SSCAL), works with any acquisition function. It admits pseudo-labels based on distances from unlabeled examples to labeled neighbors, caps their share, and reweights the loss on labeled examples to keep their expected gradient contribution unchanged. On frozen embeddings, its pseudo-labeler updates incrementally without extra model forward passes. Across two audio datasets, two frozen encoders, six acquisition functions and ten seeds, SSCAL ranks first on average among four protocols with equal training steps, answering the question affirmatively.
Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Sound (cs.SD); Audio and Speech Processing (eess.AS)
arXiv:2610.35345 [cs.LG]
(or arXiv:2610.35345v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2610.35345 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
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.
Text integral sub licența deschisă a publicației · 313 cuvinte