mlsanityData to 14 Oct 2026

Greedy Volume Maximization of Gradient Embeddings for Long-Tailed Frame-Level Bioacoustic Active Learning

arXiv 2607.13555 ↗First posted 15 Jul 2026eess.AS cs.AI cs.SDcited by 0

Not on the map. Why?

Abstract (arXiv)

Bioacoustic call-type classification relies on costly expert annotation. Active learning can reduce this burden by selecting a small batch of segments for expert annotation and using the labeled segments for training the classifier. The setting is hard: the target calls are extremely sparse and the call-type distribution is long-tailed, so a tight budget must be spent on the few rare, informative segments. We propose BADGE-Greedy-DPP, a deterministic batch selector that greedily adds the segment whose BADGE gradient embedding most enlarges the volume spanned by the batch; because this log-volume objective is submodular, the greedy rule guarantees a batch value at least a (1-1/e) fraction of the optimum of this objective, a guarantee not provided by BADGE's existing k-means++ and MCMC DPP sampling heuristics. There is also a temporal granularity mismatch in the task. The acquisition function scores whole segments, yet the informative frames inside them are few. Uniform averaging therefore washes them out. The BADGE construction naturally addresses this mismatch when applied frame-wise, as prediction residuals weight the aggregated pseudo-gradient, so confidently predicted no-call frames contribute little while a single uncertain rare-call frame can still set the segment's direction. Across 10 runs on a sparse, imbalanced hyena call-type dataset, BADGE-Greedy-DPP achieves the best overall and rare-call-type performance among all compared query strategies, including MFFT, the strongest non-BADGE baseline, and the two vanilla BADGE traversals.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 18 of its 38 references that are in the map. First posted on arXiv 15 Jul 2026.

Builds on

18 of 38 references.

References · 18

  1. 2019Class-Balanced Loss Based on Effective Number of Samplescited by 1,245
  2. 2017Deep Bayesian Active Learning with Image Datacited by 819
  3. 2019Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Boundscited by 610
  4. 2019BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learningcited by 432
  5. 2021Deep Long-Tailed Learning: A Surveycited by 417
  6. 2021Batch Active Learning at Scalecited by 127
  7. 2021Computational bioacoustics with deep learning: a review and roadmapcited by 78
  8. 2019Batch Active Learning Using Determinantal Point Processescited by 43
  9. 2020VaB-AL: Incorporating Class Imbalance and Difficulty with Variational Bayes for Active Learningcited by 25
  10. 2016Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh Distributions and Determinantal Point Processescited by 24
  11. 2021Class-Balanced Active Learning for Image Classificationcited by 15
  12. 2020Active Learning for Sound Event Detectioncited by 14
6 more
  1. 2024animal2vec and MeerKAT: A self-supervised transformer for rare-event raw audio input and a large-scale reference dataset for bioacousticscited by 14
  2. 2024Generative Active Learning for Long-tailed Instance Segmentationcited by 10
  3. 2024From Weak to Strong Sound Event Labels using Adaptive Change-Point Detection and Active Learningcited by 4
  4. 2025Aggregation Strategies for Efficient Annotation of Bioacoustic Sound Events Using Active Learningcited by 2
  5. 2025Hybrid Disagreement-Diversity Active Learning for Bioacoustic Sound Event Detectioncited by 2
  6. 2023Online Active Learning For Sound Event Detectioncited by 1

Built on by

Not cited yet.