Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Active Learning via Classifier Impact and Greedy Selection for Interactive Image Retrieval

T0 review · 3 major / 4 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read GAL is a batch-mode active learning method for interactive image retrieval that selects samples by measuring their impact on the classifier's decision boundary, outperforming existing baselines on four benchmarks and carrying a (1-1/e)…

desk verdict Solid empirical paper on active learning for interactive image retrieval with a genuine theory flaw: the (1-1/e) guarantee in §4.2.1 is unsupported and should be fixed or removed before publication. read the letter →

arxiv 2412.02310 v1 pith:4GVA3JV2 submitted 2024-12-03 cs.CV cs.IR

Leah Bar , Boaz Lerner , Nir Darshan , Rami Ben-Ari This is my paper · ORCID
classification cs.CVcs.IR
keywords activelearninginteractiveimageretrievalbatch-modegreedyselectionclassifierimpactGaussianprocesscoldstartcontent-based
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper introduces GAL, a batch-mode active-learning framework for interactive image retrieval, where each round the system shows a user a small set of images and learns from their relevance feedback. GAL's selection rule scores each unlabeled candidate by how much adding it to the training set would change the classifier, with the candidate's label guessed by the smaller of the two class-conditional changes (the pseudo-label). A greedy loop adds each chosen image to the training set before picking the next, so later choices must bring new information rather than duplicate the previous one. On four retrieval benchmarks with SVM, MLP, and Gaussian Process classifiers, GAL reports higher retrieval accuracy per labeled sample than random selection, uncertainty, diversity, and several hybrid baselines. For the Gaussian Process case the paper proves a (1-1/e) approximation guarantee for the greedy batch selection.

What carries the argument

The key machinery is a MaxMin acquisition function built on classifier impact. For SVM, each candidate $x_i$ is added to the training set once with label $+1$ and once with $-1$; the impact score is $S_i = \min_{l\in\{+1,-1\}} \|W(x_i,l) - W_0\|^2$, where $W_0$ is the current hyperplane and $W(x_i,l)$ is the hyperplane retrained with the candidate added (the minimizing label is the pseudo-label). For the MLP, the same idea applies to the vector of all network weights. For the Gaussian Process, the acquisition function is $F_{\rm gp}(x_i) = -\big(\sum_{x\in X_c} \sigma^2_{A\cup\{x_i\}}(x) + \alpha \max_{x\in X_c} \sigma^2_{A\cup\{x_i\}}(x)\big)$, i.e., the negative total and worst-case predictive variance after adding the candidate to the training set $A$; for the GP the variance reduction is label-independent, so no pseudo-label is needed. The greedy Algorithm 1 then inserts each selected point (with its pseudo-label) into the training set before selecting the next, which yields a conditioned diversity: nearby points have low impact once a similar point is in the batch, while boundary points remain attractive.

What would settle it

Measure how often GAL's pseudo-labels match the actual user labels for the images it selects over the first few active-learning cycles on a real retrieval dataset. If that agreement is near chance (50%) while GAL still beats random and diversity baselines, then the minimal-shift pseudo-label mechanism is not the source of the gains, and the paper's explanation of its own success fails.

Watch

Extended reading notes

Core claim

The core claim is that in the cold-start, open-set, class-imbalanced setting of interactive image retrieval, the most useful unlabeled samples are those that would most strongly reshape the classifier's decision boundary, and that this can be estimated without oracle labels. For SVM and MLP classifiers, GAL tentatively labels each candidate as positive and as negative, computes the resulting change in the decision boundary (or in all network weights for the MLP), and treats the label producing the smaller change as the pseudo-label, since a wrong label would move the boundary more. The impact score of the candidate is that smaller change, and the algorithm selects the candidate with the largest score. For Gaussian Processes, the acquisition function instead measures how much adding the candidate would reduce the total and worst-case predictive variance over the candidate pool. The greedy loop then iterates: after a point is selected it is added to the training set with its pseudo-label, so the next pick cannot simply echo the information already gained. The paper shows this mechanism outperforms existing active-learning baselines across Paris, Places, FSOD-IR, and MIRFLICKR, and proves that the GP acquisition function is monotone submodular, so greedy batch selection inherits the classical (1-1/e) approximation bound.

Load-bearing premise

The method's sample selection only works if, for each candidate image, the label that causes the smallest change in the trained classifier is usually the correct label—the pseudo-label—because the impact scores and the whole greedy batch construction are built on those guessed labels.

Editorial extensions

If this is right

  • An interactive retrieval system can reach a target mean average precision with fewer user-labeled images per round, since GAL selects images that most quickly sharpen the classifier.
  • The greedy construction makes larger batch sizes usable in cold-start settings; the paper's experiments show GAL stays ahead of baselines even when the batch budget grows from 3 to 7 after the first cycles.
  • For Gaussian-Process retrieval, the selected batch is provably near-optimal with respect to uncertainty reduction, by the (1 - 1/e) submodular maximization guarantee.
  • Restricting the selection pool to the top-K ranked candidates (e.g., K=200) improves both accuracy and runtime, so the method scales to databases where scoring every image each round is too slow.
  • The newly introduced FSOD-IR benchmark offers an object-level interactive retrieval testbed where queries are objects inside images, which is more challenging than image-level retrieval.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The minimal-shift pseudo-label heuristic could transfer to other severely imbalanced binary active-learning settings, such as defect detection or medical triage; there it could be tested by comparing pseudo-label accuracy with oracle labels on real workloads.
  • The paper's evidence that a small top-K pool beats the full pool suggests that the optimal pool size depends on feature-space geometry; a natural follow-up is to measure how K should scale with database size and class rarity.
  • The (1 - 1/e) guarantee covers only the GP acquisition function; for SVM and MLP the greedy scheme has no formal bound, so an open question is whether a submodular relaxation or a different proof route could extend the guarantee to those settings.
  • Because the pseudo-label accuracy is initially modest (the paper reports it hovering around chance-to-60% at the start on FSOD), combining GAL with a first cycle of random or diversity-based selection could improve cold-start robustness even further.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes GAL, a batch-mode active learning framework for interactive content-based image retrieval. The main idea is to select samples by their estimated impact on the classifier: for SVM and MLP classifiers this impact is the change in the decision boundary under a pseudo-label chosen to minimize that change (Eqs. 3-6), while for Gaussian processes it is the reduction of global predictive variance, with a greedy batch construction that adds each pseudo-labeled selection to the training set before choosing the next sample. The paper reports experiments on Paris-6K, Places365, a new FSOD-IR object-level benchmark, and MIRFLICKR-25K, comparing against random selection, RBMAL, COD, MaxiMin, Kmeans++, Coreset, ITAL, and Kapoor et al. A theoretical claim is made that greedy GP-based selection satisfies a (1-1/e) approximation guarantee via the Nemhauser et al. theorem. Code is available online.

Significance. If the empirical claims hold, GAL is a useful contribution to interactive image retrieval active learning: it addresses cold-start, open-set, and class-imbalance conditions with a unified acquisition framework, introduces a new object-level benchmark (FSOD-IR), and provides code. The breadth of the evaluation (three classifier families, four datasets, several baselines, and a batch-mode ablation) is a clear strength. However, the theoretical guarantee stated as contribution 3 is not established by the current proof, and the pseudo-label proxy underlying the SVM/MLP acquisition functions is not validated on real data. The empirical comparison would be stronger with variance reporting or significance tests.

major comments (3)
  1. [§4.2.1, Eq. (15)] The claimed (1-1/e) guarantee is not established. For the squared-exponential kernel in Eq. (10), σ²_∅(x)=K(x,x)=1, so F(∅)=-(|Xc|+α)<0, violating the non-negativity condition of the Nemhauser et al. theorem. More importantly, the submodularity argument is not valid as written: Eq. (15) contains a negative sign and a max term, and while the sum Σ_x σ²_A(x) is monotone submodular by Das & Kempe (2008), a maximum of monotone submodular functions is not submodular in general, and no GP-specific property is shown for the max term. A constant shift would restore non-negativity but would not fix the submodularity issue. Therefore contribution 3, as stated, is unsupported.
  2. [§3, Eq. (3)] The pseudo-label proxy is load-bearing for the SVM and MLP acquisition functions: Eq. (3) assumes the true label produces a smaller decision-boundary shift than the false label, so the argmin label approximates the true label. The paper motivates this only with a 2D Gaussian toy example (Fig. 2), and the pseudo-label accuracy reported in Fig. 11b covers only FSOD-IR with B=1 and one candidate size. Since Algorithm 1 inserts each selected sample with its pseudo-label into the training set before choosing the next sample, incorrect pseudo-labels can systematically steer within-batch selections. Please validate this assumption on the real datasets, e.g., by reporting pseudo-label accuracy per dataset/cycle or by ablating with oracle labels during greedy selection.
  3. [Tables 2-6 and Figs. 8-14] The empirical comparisons are averaged over only five random queries, but no standard deviations, confidence intervals, or significance tests are reported. Some observed margins are small (e.g., Table 2, Paris K=200: GAL 0.960 vs Random 0.942; Table 3, Paris K=100: GAL 0.946 vs Kmeans++ 0.922), and with N=5 these differences may not be reliable. The "superiority" claim would be much stronger with per-query variability or paired statistical tests over the same query set.
minor comments (4)
  1. [Algorithm 1] For the GP acquisition function, the pseudo-label returned by Next is Null, but Algorithm 1 still performs Yl ← Yl ∪ {l*}, which would insert Null into the label set used for subsequent training; the algorithm should make explicit that GP selection does not update Yl with pseudo-labels.
  2. [§4.2] There is an overloaded use of F: Eq. (13) defines Fgp as a function of a candidate point xi, while Eq. (15) defines F(A) as a set function over a batch A; please use distinct symbols or clarify the relationship.
  3. [§4.1, Eq. (7)] Eq. (7) defines Fmlp as a norm but does not specify whether it is the L2 norm or a squared norm; the SVM version in Eq. (1) uses squared L2, so the MLP definition should be made consistent.
  4. [Throughout] There are several typos, e.g., "aqcusition function" in §4.2, "lake size" in §5.1, and "Kappor et al." in Table 6; these should be corrected.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: GAL defines acquisition functions from classifier behavior, evaluates against external baselines, and its only self-citation is a non-load-bearing related-work pointer.

full rationale

The paper's derivation chain is self-contained. The SVM/MLP acquisition functions (Eqs. 1, 4, 6, 7) are defined as the norm of the change in classifier parameters when a candidate point is added with each possible label, and the pseudo-label is the minimizer of that change (Eq. 3); no parameter of these functions is fitted to the retrieval metric, so the selection scores are not equivalent by construction to the reported mAP gains. The GP acquisition (Eq. 13) directly minimizes the posterior variance of the candidate set and requires no labels, and the greedy batch construction is an explicit algorithmic procedure whose effect is measured against GAL(batch) and external methods. Empirical support comes from comparisons to Random, RBMAL, COD, Kmeans++, Coreset, MaxiMin, ITAL, and Kapoor et al., which are independent of the authors' prior work. The paper itself flags dependence on pseudo-label reliability in Section 4.1.1 (Eq. 8 and Fig. 6), which is an assumption/limitation rather than a circular step. The only self-citation, Lerner et al. (2023), appears in a list of IIR works in the introduction and is not used to justify any load-bearing claim. The theoretical guarantee in Section 4.2.1 is open to challenge on mathematical grounds (e.g., F(empty) is negative as written, and a maximum of submodular functions need not be submodular), but that is a correctness risk, not circularity: the proposed bound is not built into the definition of the acquisition function or imported from the authors' own prior results.

Assumptions & free parameters 5 free parameters · 3 assumptions · 0 invented entities

The central method is a heuristic acquisition function combined with a greedy batch scheme. The main free parameters are hyperparameters (alpha, K, kernel length scale, SVM C) that are either unreported or varied across experiments. The pseudo-label proxy is the key ad hoc assumption that everything downstream depends on.

free parameters (5)
  • alpha
    Weight of the L-infinity term in the GP acquisition function (Eq. 13). No value or tuning procedure is reported.
  • candidate pool size K = 100, 200, 1k, all (varies by experiment)
    Hyperparameter controlling the candidate subset. The paper suggests a heuristic (a few percent of the dataset) but reports results across several values and selects K=200 for main comparisons.
  • GP kernel length scale gamma
    Appears in the squared exponential kernel (Eq. 10). No value or tuning method is specified.
  • SVM regularization C
    Not specified anywhere in the experimental setup.
  • MLP hyperparameters = 3 layers, 10 inner neurons (architecture only)
    Number of epochs, learning rate, optimizer, and weight initialization are not reported.
assumptions (3)
  • ad hoc to paper Pseudo-label proxy: for a candidate x_i, the label that minimizes the decision-boundary shift approximates the true label.
    Introduced in Section 3 (after Fig. 2) and formalized in Eq. (3). The entire acquisition function S_i depends on this; no proof or real-data validation is provided.
  • domain assumption Pretrained ImageNet features are transferable and sufficient for IIR classification.
    Section 4 states features come from a pretrained backbone (ResNet-50 on ImageNet-21K). If these features do not separate query classes, both the classifier and the acquisition functions degrade.
  • standard math GP predictive variance reduction is monotone and submodular (Das & Kempe, 2008).
    Used in Section 4.2.1 to claim the acquisition function (15) is submodular. The cited result holds for variance reduction for each test point, but the extension to the maximum term in (15) is not justified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Active Learning via Classifier Impact and Greedy Selection for Interactive Image Retrieval." pith.science (2026). https://pith.science/paper/4GVA3JV2

@misc{pith2026241202310,
  author       = {Pith},
  title        = {Pith review of: Active Learning via Classifier Impact and Greedy Selection for Interactive Image Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4GVA3JV2}},
  note         = {Machine review of arXiv:2412.02310}
}
read the original abstract

Active Learning (AL) is a user-interactive approach aimed at reducing annotation costs by selecting the most crucial examples to label. Although AL has been extensively studied for image classification tasks, the specific scenario of interactive image retrieval has received relatively little attention. This scenario presents unique characteristics, including an open-set and class-imbalanced binary classification, starting with very few labeled samples. We introduce a novel batch-mode Active Learning framework named GAL (Greedy Active Learning) that better copes with this application. It incorporates a new acquisition function for sample selection that measures the impact of each unlabeled sample on the classifier. We further embed this strategy in a greedy selection approach, better exploiting the samples within each batch. We evaluate our framework with both linear (SVM) and non-linear MLP/Gaussian Process classifiers. For the Gaussian Process case, we show a theoretical guarantee on the greedy approximation. Finally, we assess our performance for the interactive content-based image retrieval task on several benchmarks and demonstrate its superiority over existing approaches and common baselines. Code is available at https://github.com/barleah/GreedyAL.

Figures

Figures reproduced from arXiv: 2412.02310 by the authors.

Figure 1
Figure 1. Main flow of the AL cycle. The top-K candidate set at cycle [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Label proxy demonstration: The points are sampled from two Gaussian distributions, demonstrat [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. In a 2D Gaussian toy example, we illustrate a binary class scenario characterized by an imbalanced [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (11 more)
Figure 4
Figure 4. Figure 4: To calculate the score for a point xi in the candidate set, we train a classifier C(θ + i ) by assuming the sample is positive. Similarly, we train another classifier C(θ − i ) with a negative label. The impact value Si is then determined as the minimum value obtained …
Figure 5
Figure 5. Figure 5: In the SVM scenario, the GAL algorithm employs a binary tree structure. The initial point [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: Theoretical results for the normalized probability of obtaining [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Gaussian Process: The true function is represented by a dashed blue line, while the prediction [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: mAP Learning Curves of SVM-based GAL with [PITH_FULL_IMAGE:figures/full_fig_p017_8.png]
Figure 9
Figure 9. Figure 9: mAP Learning Curves of SVM-based GAL with [PITH_FULL_IMAGE:figures/full_fig_p018_9.png]
Figure 10
Figure 10. Figure 10: mAP Learning Curves of SVM-based GAL with [PITH_FULL_IMAGE:figures/full_fig_p018_10.png]
Figure 11
Figure 11. Figure 11: (a) mAP Learning Curves of SVM-based GAL with [PITH_FULL_IMAGE:figures/full_fig_p019_11.png]
Figure 12
Figure 12. Figure 12: Image retrieval results for Tin Can in FSOD-IR dataset with B = 3 at iteration 4. Green boxes stand for relevant results while red boxes account for false positives. The second query image has two objects: Can and Display monitor. The RBMAL method mistakenly retrieves…
Figure 13
Figure 13. Figure 13: (a) mAP Learning Curves of MLP-based AL selection with [PITH_FULL_IMAGE:figures/full_fig_p021_13.png]
Figure 14
Figure 14. Figure 14: mAP Learning Curves of GP-based GAL with [PITH_FULL_IMAGE:figures/full_fig_p021_14.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Foundation-Assisted Active Learning for Object Detection Annotation

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A foundation-model-assisted active learning system using UPN+SAM2 reference boxes and DINOv2 object features improves low-budget annotation efficiency for remote sensing object detection.

Reference graph

Works this paper leans on

60 extracted references · 55 canonical work pages · cited by 1 Pith paper

  1. [1]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 global.max substring 't := if while FUNCTION format.date year duplicate empty "emp...

  2. [2]

    Contextual diversity for active learning

    Sharat Agarwal, Himanshu Arora, Saket Anand, and Chetan" Arora. Contextual diversity for active learning. In ECCV, 2020

    work page 2020

  3. [3]

    Active learning for imbalanced datasets

    Umang Aggarwal, Adrian Popescu, and Celine Hudelot. Active learning for imbalanced datasets . WACV 2020, 2 0 (1): 0 1417--1426, 2020

    work page 2020

  4. [4]

    Deep batch active learning by diverse, uncertain gradient lower bounds

    Jordan T Ash, Chicheng Zhang, Akshay Krishnamurthy, John Langford, and Alekh Agarwal. Deep batch active learning by diverse, uncertain gradient lower bounds. ICLR, 2020

    work page 2020

  5. [5]

    Semi-supervised few-shot learning with prototypical random walk

    Ahmed Ayyad, Li Yuchen, Muaz Raden, Albarqouni Shadi, and Mohamed Elhoseiny. Semi-supervised few-shot learning with prototypical random walk. In AAAI Workshop on Meta-Learning and MetaDL Challenge, 2021

    work page 2021

  6. [6]

    Relevance feedback for enhancing content based image retrieval and automatic prediction of semantic image features: Application to bone tumor radiographs

    Imon Banerjee, Camille Kurtz, Alon Edward Devorah, Bao Do, Daniel L Rubin, and Christopher F Beaulieu. Relevance feedback for enhancing content based image retrieval and automatic prediction of semantic image features: Application to bone tumor radiographs. Journal of biomedical informatics, 84: 0 123--135, 2018

    work page 2018

  7. [7]

    Content-based image retrieval and the semantic gap in the deep learning era

    Bj \"o rn Barz and Joachim Denzler. Content-based image retrieval and the semantic gap in the deep learning era. In ICPR, pp.\ 245--260, 2021

    work page 2021

  8. [8]

    o rn Barz, Christoph K \

    Bj \"o rn Barz, Christoph K \"a ding, and Joachim Denzler. Information-theoretic active learning for content-based image retrieval. In German Conference on Pattern Recognition, pp.\ 650--666. Springer, 2018

    work page 2018

Show all 60 references
  1. [9]

    Incorporating diversity in active learning with support vector machines

    Klaus Brinker. Incorporating diversity in active learning with support vector machines. In ICML, pp.\ 59--66, 2003

    2003

  2. [10]

    Ranked batch-mode active learning

    Thiago NC Cardoso, Rodrigo M Silva, S \'e rgio Canuto, Mirella M Moro, and Marcos A Gon c alves. Ranked batch-mode active learning. Information Sciences, 379: 0 313--337, 2017

    2017

  3. [11]

    On initial pools for deep active learning

    Akshay L Chandra, Sai Vikas Desai, Chaitanya Devaguptapu, and Vineeth N Balasubramanian. On initial pools for deep active learning. In NeurIPS 2020 Workshop on Pre-registration in Machine Learning, pp.\ 14--32. PMLR, 2021

    2020

  4. [12]

    Chapelle

    O. Chapelle. Training a support vector machine in the primal. Neural Computation, 19 0 (5): 0 1155--1178, 2007

    2007

  5. [13]

    Batch Active Learning at Scale

    Gui Citovsky, Giulia DeSalvo, Claudio Gentile, Lazaros Karydas, Anand Rajagopalan, Afshin Rostamizadeh, and Sanjiv Kumar. Batch Active Learning at Scale . NeurIPS, 2021

    2021

  6. [14]

    Das and D

    A. Das and D. Kempe. Algorithms for subset selection in linear regression. In ACM Symposium on the Theory of Computing (STOC), pp.\ 45--54, 2008

    2008

  7. [15]

    Active Learning at the ImageNet Scale

    Zeyad Emam, Sami Ali, Chu Hong-Min, Chiang Ping-Yeh, Czaja Wojciech, Leapman Richard, Goldblum Micah, and Goldstein Tom. Active Learning at the ImageNet Scale . In arXiv:2111.12880, 2021

    2021 arXiv

  8. [16]

    Few-Shot Object Detection with Attention-RPN and Multi-Relation Detector

    Qi Fan, Wei Zhuo, Chi-Keung Tang, and Yu-Wing Tai. Few-Shot Object Detection with Attention-RPN and Multi-Relation Detector . In CVPR, 2020

    2020

  9. [17]

    Submodular Functions and Optimization

    Satoru Fujishige. Submodular Functions and Optimization. Elsevier, 2005

    2005

  10. [18]

    Dropout as a bayesian approximation: Representing model uncertainty in deep learning

    Yarin Gal and Zoubin Ghahramani. Dropout as a bayesian approximation: Representing model uncertainty in deep learning. ICML, 2016

    2016

  11. [19]

    Consistency-based semi-supervised active learning: Towards minimizing labeling cost

    Mingfei Gao, Zizhao Zhang, Guo Yu, Sercan \"O Ar k, Larry S Davis, and Tomas Pfister. Consistency-based semi-supervised active learning: Towards minimizing labeling cost. In ECCV, 2020

    2020

  12. [20]

    Meta agent teaming active learning for pose estimation

    Jia Gong, Zhipeng Fan, Qiuhong Ke, Hossein Rahmani, and Jun Liu. Meta agent teaming active learning for pose estimation. In CVPR, 2022

    2022

  13. [21]

    Active learning methods for interactive image retrieval

    Philippe Henri Gosselin and Matthieu Cord. Active learning methods for interactive image retrieval . IEEE Transactions on Image Processing, 17 0 (7): 0 1200--1211, 2008

    2008

  14. [22]

    Active learning on a budget: Opposite strategies suit high and low budgets

    Guy Hacohen, Avihu Dekel, and Daphna Weinshall. Active learning on a budget: Opposite strategies suit high and low budgets. ICML, 2022

    2022

  15. [23]

    Semi-supervised active learning with temporal output discrepancy

    Siyu Huang, Tianyang Wang, Haoyi Xiong, Jun Huan, and Dejing Dou. Semi-supervised active learning with temporal output discrepancy. In ICCV, 2021

    2021

  16. [24]

    The MIRFLICKR retrieval evaluation

    Mark J Huiskes and Michael S Lew. The MIRFLICKR retrieval evaluation. In Proceedings of the 1st ACM international conference on Multimedia information retrieval, pp.\ 39--43, 2008

    2008

  17. [25]

    Cold-start active learning for image classification

    Qiuye Jin, Mingzhi Yuan, Shiman Li, Haoran Wang, Manning Wang, and Zhijian Song. Cold-start active learning for image classification. Information Sciences, 616: 0 16--36, 2022

    2022

  18. [26]

    Kapoor, K

    A. Kapoor, K. Grauman, R. Urtasun, and T. Darrell. Active learning with gaussian processes for object categorization. In ICCV, 2007

    2007

  19. [27]

    Mina Karzand and Robert D. Nowak. Maximin active learning in overparameterized model classes. IEEE Journal on Selected Areas in Information Theory, 1 0 (1): 0 167--177, 2020

    2020

  20. [28]

    V. Khakham. K-segment. https://github.com/vkhakham/k-segment, 2019. GitHub repository

    2019

  21. [29]

    SIMILAR : Submodular information measures based active learning in realistic scenarios

    Suraj Kothawade, Nathan Beck, Krishnateja Killamsetty, and Rishabh Iyer. SIMILAR : Submodular information measures based active learning in realistic scenarios. NeurIPS, 2021

    2021

  22. [30]

    Advancing image retrieval with few-shot learning and relevance feedback

    Boaz Lerner, Nir Darshan, and Rami Ben-Ari. Advancing image retrieval with few-shot learning and relevance feedback. arXiv:2312.11078, 2023

    2023 arXiv

  23. [31]

    Deep reinforcement active learning for human-in-the-loop person re-identification

    Zimo Liu, Jingya Wang, Shaogang Gong, Dacheng Tao, and Huchuan Lu. Deep reinforcement active learning for human-in-the-loop person re-identification. In ICCV, 2019

    2019

  24. [32]

    Manjunath, P

    B.S. Manjunath, P. Wu, S. Newsam, and H.D. Shin. A texture descriptor for browsing and similarity retrieval. Signal Processing: Image Communication, 2000

    2000

  25. [33]

    Fast Interactive Image Retrieval using large-scale unlabeled data

    Akshay Mehra, Jihun Hamm, and Mikhail Belkin. Fast Interactive Image Retrieval using large-scale unlabeled data . KDD, 2018

    2018

  26. [34]

    Nemhauser, L.A

    G.L. Nemhauser, L.A. Wolsey, and M. L. Fisher. An analysis of approximations for maximizing submodular set functions- I . Mathematical Programming, 14: 0 265--294, 1978

    1978

  27. [35]

    Image Retrieval with relevance feedback using SVM active learning

    Giang Truong Ngo, Tao Quoc Ngo, and Dung Duc Nguyen. Image Retrieval with relevance feedback using SVM active learning . International Journal of Electrical and Computer Engineering, 6 0 (6): 0 3238--3246, 2016

    2016

  28. [36]

    Deep neural networks are easily fooled: High confidence predictions for unrecognizable images

    Anh Nguyen, Jason Yosinski, and Jeff Clune. Deep neural networks are easily fooled: High confidence predictions for unrecognizable images. CVPR, 2015

    2015

  29. [37]

    Active learning for open-set annotation

    Kun-Peng Ning, Xun Zhao, Yu Li, and Sheng-Jun Huang. Active learning for open-set annotation. In CVPR, 2022

    2022

  30. [38]

    Active learning by feature mixing

    Amin Parvaneh, Ehsan Abbasnejad, Damien Teney, Gholamreza Reza Haffari, Anton van den Hengel, and Javen Qinfeng Shi. Active learning by feature mixing. In CVPR, 2022

    2022

  31. [39]

    On the utility of active instance selection for few-shot learning

    Pouya Pezeshkpour, Zhengli Zhao, and Sameer Singh. On the utility of active instance selection for few-shot learning. NeurIPS HAMLETS, 2020

    2020

  32. [40]

    Convolutional neural networks for relevance feedback in content based image retrieval

    Lorenzo Putzu, Luca Piras, and Giorgio Giacinto. Convolutional neural networks for relevance feedback in content based image retrieval. Multimedia Tools and Applications, 79: 0 26995--27021, 2020

    2020

  33. [41]

    A short-term learning framework based on relevance feedback for content-based image retrieval

    Hamed Qazanfari, Hamid Hassanpour, and Kazem Qazanfari. A short-term learning framework based on relevance feedback for content-based image retrieval. In 2017 3rd Iranian Conference on Intelligent Systems and Signal Processing (ICSPIS), pp.\ 136--140. IEEE, 2017

    2017

  34. [42]

    Revisiting O xford and P aris: Large-scale image retrieval benchmarking

    Filip Radenovi \'c , Ahmet Iscen, Giorgos Tolias, Yannis Avrithis, and Ond r ej Chum. Revisiting O xford and P aris: Large-scale image retrieval benchmarking. In CVPR, 2018

    2018

  35. [43]

    A novel relevance feedback method for CBIR

    Yunbo Rao, Wei Liu, Bojiang Fan, Jiali Song, and Yang Yang. A novel relevance feedback method for CBIR . World Wide Web, 21: 0 1505--1522, 2018

    2018

  36. [44]

    Imagenet-21k pretraining for the masses

    Tal Ridnik, Emanuel Ben-Baruch, Asaf Noy, and Lihi Zelnik-Manor. Imagenet-21k pretraining for the masses. NeurIPS Datasets and Benchmarks Track , 2021

    2021

  37. [45]

    Instance-based relevance feedback for image retrieval

    Fabio Roli et al. Instance-based relevance feedback for image retrieval. Advances in neural information processing systems, 17, 2004

    2004

  38. [46]

    Active learning for convolutional neural networks: A Core-Set approach

    Ozan Sener and Silvio Savarese. Active learning for convolutional neural networks: A Core-Set approach . ICML, 2018

    2018

  39. [47]

    Support vector machine active learning for image retrieval

    Simon Tong and Edward Chang. Support vector machine active learning for image retrieval. In Proceedings of the ninth ACM international conference on Multimedia, pp.\ 107--118, 2001

    2001

  40. [48]

    Support vector machine active learning with applications to text classification

    Simon Tong and Daphne Koller. Support vector machine active learning with applications to text classification. Journal of machine learning research, 2 0 (Nov): 0 45--66, 2001

    2001

  41. [49]

    k-means++: The advantages of careful seeding

    Sergei Vassilvitskii and David Arthur. k-means++: The advantages of careful seeding. In ACM-SIAM symposium on Discrete algorithms, pp.\ 1027--1035, 2006

    2006

  42. [50]

    Cost-effective active learning for deep image classification

    Keze Wang, Dongyu Zhang, Ya Li, Ruimao Zhang, and Liang Lin. Cost-effective active learning for deep image classification. IEEE Transactions on Circuits and Systems for Video Technology, 27 0 (12): 0 2591--2600, 2016

    2016

  43. [51]

    Boosting active learning via improving test performance

    Tianyang Wang, Xingjian Li, Pengkun Yang, Guosheng Hu, Xiangrui Zeng, Siyu Huang, Cheng-Zhong Xu, and Min Xu. Boosting active learning via improving test performance. In Proceedings of the AAAI Conference on Artificial Intelligence, pp.\ 8566--8574, 2022

    2022

  44. [52]

    Gaussian processes for machine learning, volume 2

    Christopher KI Williams and Carl Edward Rasmussen. Gaussian processes for machine learning, volume 2. MIT press Cambridge, MA, 2006

    2006

  45. [53]

    Entropy-based active learning for object detection with progressive diversity constraint

    Jiaxi Wu, Jiaxin Chen, and Di Huang. Entropy-based active learning for object detection with progressive diversity constraint. In CVPR, 2022

    2022

  46. [54]

    Towards general and efficient active learning

    Yichen Xie, Tomizuka Masayoshi, and Zhan Wei. Towards general and efficient active learning. In arXiv:2112.07963, 2021

    2021 arXiv

  47. [55]

    Multi-class active learning by uncertainty sampling with diversity maximization

    Yi Yang, Zhigang Ma, Feiping Nie, Xiaojun Chang, and Alexander G Hauptmann. Multi-class active learning by uncertainty sampling with diversity maximization. International Journal of Computer Vision, 113 0 (2): 0 113--127, 2015

    2015

  48. [56]

    Active learning through a covering lens

    Ofer Yehuda, Avihu Dekel, Guy Hacohen, and Daphna Weinshall. Active learning through a covering lens. NeurIPS, 2022

    2022

  49. [57]

    Cold-start active learning through self-supervised language modeling

    Michelle Yuan, Hsuan-Tien Lin, and Jordan Boyd-Graber. Cold-start active learning through self-supervised language modeling. arXiv:2010.09535, 2020

    2010 arXiv

  50. [58]

    Zhang, S

    Q. Zhang, S. A. Goldman, W. Yu, and J. Fritts. Content-based image retrieval using multiple-instance learning. In ICML, 2002

    2002

  51. [59]

    G. Zhao, E. Dougherty, B.J. Yoon, F. Alexander, and X. Qian. Efficient active learning for gaussian process classification by error reduction. In NeurIPS , 2021

    2021

  52. [60]

    Places: A 10 million image database for scene recognition

    Bolei Zhou, Agata Lapedriza, Aditya Khosla, Aude Oliva, and Antonio Torralba. Places: A 10 million image database for scene recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 40 0 (6): 0 1452--1464, 2017

    2017

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.