mlsanityData to 7 Oct 2026

Stochastic Batch Acquisition: A Simple Baseline for Deep Active Learning

arXiv 2106.12059 ↗First posted 22 Jun 2021cs.LG stat.MLcited by 36

Not on the map. Why?

Abstract (arXiv)

We examine a simple stochastic strategy for adapting well-known single-point acquisition functions to allow batch active learning. Unlike acquiring the top-K points from the pool set, score- or rank-based sampling takes into account that acquisition scores change as new data are acquired. This simple strategy for adapting standard single-sample acquisition strategies can even perform just as well as compute-intensive state-of-the-art batch acquisition functions, like BatchBALD or BADGE, while using orders of magnitude less compute. In addition to providing a practical option for machine learning practitioners, the surprising success of the proposed method in a wide range of experimental settings raises a difficult question for the field: when are these expensive batch acquisition methods pulling their weight?

Key facts. Cited by 36 papers in the map (data through 7 Oct 2026). Builds on 20 of its 66 references that are in the map. First posted on arXiv 22 Jun 2021.

Builds on

20 of 66 references.

References · 20

  1. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  2. 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
  3. 2015Prioritized Experience Replaycited by 1,070
  4. 2017Deep Bayesian Active Learning with Image Datacited by 812
  5. 2019Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Boundscited by 602
  6. 2019BatchBALD: Efficient and Diverse Batch Acquisition for Deep Bayesian Active Learningcited by 428
  7. 2019An Evaluation Dataset for Intent Classification and Out-of-Scope Predictioncited by 378
  8. 2017EMNIST: an extension of MNIST to handwritten letterscited by 212
  9. 2019Stochastic Beams and Where to Find Them: The Gumbel-Top-k Trick for Sampling Sequences Without Replacementcited by 205
  10. 2021Deep Deterministic Uncertainty: A Simple Baselinecited by 200
  11. 2014A* Samplingcited by 150
  12. 2021Gone Fishing: Neural Active Learning with Fisher Embeddingscited by 88
8 more
  1. 2021On Statistical Bias In Active Learning: How and When To Fix Itcited by 72
  2. 2017Boltzmann Exploration Done Rightcited by 69
  3. 2019Bayesian Batch Active Learning as Sparse Subset Approximationcited by 68
  4. 2019Batch Active Learning Using Determinantal Point Processescited by 42
  5. 2020Bayesian active learning for production, a systematic study and a reusable librarycited by 15
  6. 2021Can Active Learning Preemptively Mitigate Fairness Issues?cited by 15
  7. 2020Synbols: Probing Learning Algorithms with Synthetic Datasetscited by 9
  8. 2021Low-Regret Active learningcited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 2; 2022: 3; 2023: 3; 2024: 10; 2025: 11; 2026: 7.

36 papers.

Sort:

  1. 2024Rethinking Aleatoric and Epistemic Uncertaintycited by 40
  2. 2022Unifying Approaches in Active Learning and Active Sampling via Fisher Information and Information-Theoretic Quantitiescited by 28
  3. 2021Causal-BALD: Deep Bayesian Active Learning of Outcomes to Infer Treatment-Effects from Observational Datacited by 27
  4. 2024Active Learning for Neural PDE Solverscited by 22
  5. 2024Deep Bayesian Active Learning for Preference Modeling in Large Language Modelscited by 19
  6. 2023Multi-Fidelity Active Learning with GFlowNetscited by 15
  7. 2023Black-Box Batch Active Learning for Regressioncited by 12
  8. 2024Making Better Use of Unlabelled Data in Bayesian Active Learningcited by 10
  9. 2024Querying Easily Flip-flopped Samples for Deep Active Learningcited by 7
  10. 2025Scaling Textual Gradients via Sampling-Based Momentumcited by 5
  11. 2025Active Learning with Selective Time-Step Acquisition for PDEscited by 4
  12. 2025Fusion or Confusion? Multimodal Complexity Is Not All You Needcited by 4
24 more
  1. 2024Progressive Generalization Risk Reduction for Data-Efficient Causal Effect Estimationcited by 3
  2. 2025Enhancing Treatment Effect Estimation via Active Learning: A Counterfactual Covering Perspectivecited by 3
  3. 2021Depth Uncertainty Networks for Active Learningcited by 2
  4. 2022PyRelationAL: a python library for active learning research and developmentcited by 2
  5. 2022Efficient Bayesian Updates for Deep Active Learning via Laplace Approximationscited by 2
  6. 2023Speeding Up BatchBALD: A k-BALD Family of Approximations for Active Learningcited by 2
  7. 2025Causal-EPIG: A Prediction-Oriented Active Learning Framework for CATE Estimationcited by 2
  8. 2024Information-Theoretic Active Correlation Clusteringcited by 1
  9. 2024Active Learning to Guide Labeling Efforts for Question Difficulty Estimationcited by 1
  10. 2024MALADY: Multiclass Active Learning with Auction Dynamics on Graphscited by 1
  11. 2024Learning Set Functions with Implicit Differentiationcited by 1
  12. 2025ActiveCQ: Active Estimation of Causal Quantitiescited by 1
  13. 2025Towards Open-Ended Discovery for Low-Resource NLPcited by 1
  14. 2026Finally Outshining the Random Baseline: A Simple and Effective Solution for Active Learning in 3D Biomedical Imagingcited by 1
  15. 2026Observationally Informed Adaptive Causal Experimental Designcited by 1
  16. 2026BoSS: A Best-of-Strategies Selector as an Oracle for Deep Active Learningcited by 1
  17. 2026Learning Where to Simulate: Generative Active Sampling for Online PDE Surrogate Trainingcited by 1
  18. 2025ConBatch-BAL: Batch Bayesian Active Learning under Budget Constraintscited by 0
  19. 2025Batch Bayesian Active Learning with Partial Batch Label Samplingcited by 0
  20. 2025Diffusion-Driven Two-Stage Active Learning for Low-Budget Semantic Segmentationcited by 0
  21. 2025Active Learning with Task-Driven Representations for Messy Poolscited by 0
  22. 2026Active Learning Using Aggregated Acquisition Functions: Accuracy and Sustainability Analysiscited by 0
  23. 2026Conformal Cross-Modal Active Learningcited by 0
  24. 2026Data-Efficient Neural Operator Training via Physics-Based Active Learningcited by 0