Pith. sign in

REVIEW 3 major objections 6 minor 51 references

Know2Vec: A Black-Box Proxy for Neural Network Retrieval

T0 review · 3 major / 6 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 Know2Vec claims a black-box proxy can vectorize neural network knowledge from random probe samples and align it with query tasks, achieving superior retrieval accuracy while preserving model privacy.

desk verdict Know2Vec is an empirically strong black-box model retrieval pipeline whose theoretical justification (Lemma 1) does not hold up; it deserves peer review with a demanded fix. read the letter →

arxiv 2412.16251 v1 pith:L5LPCFYD submitted 2024-12-20 cs.LG

Zhuoyi Shang , Yanwei Liu , Jinxia Liu , Xiaoyan Gu , Ying Ding , Xiangyang Ji This is my paper · ORCID
classification cs.LG
keywords neuralnetworkretrievalblack-boxproxymodelvectorizationknowledgeconsistencyrepresentationmatrixdecisionboundarysamplessource-freetransferabilityestimationzoo
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 proposes Know2Vec, a black-box proxy for retrieving the most suitable pre-trained neural network for a given query task without access to model parameters or training data. It aims to prove that a model's decision knowledge can be captured through its responses to probe samples, encoded into vectors, and aligned with query-task vectors in a learned embedding space. If true, model-marketplace users could find a good model by submitting only a small labeled dataset, while model owners keep weights and training data private. The paper reports top-1 retrieval accuracy of 94.82%, outperforming statistical, dynamic, and LLM-based baselines, and offers a theoretical lemma (Lemma 1) claiming random probe datasets can substitute for the model's own training centroids.

What carries the argument

The load-bearing object is the Knowledge Representation Matrix (KRM), whose rows are perturbation vectors r_b^a = x_b^a - x_a between class centroid x_a and decision-boundary sample x_b^a. Lemma 1 is the mechanism that lets Know2Vec use arbitrary probe datasets instead of training data: it claims probes with decision values near plus or minus one reconstruct the KRM offsets up to small $\sigma$ terms. The KRM is then organized into a directed graph set G_Phi per class, and an inner-outer bidirectional LSTM encoder converts G_Phi into the model vector h, while a class-mean LSTM converts the query into t. Cosine similarity with a margin of 0.4 in a supervised alignment space finally ranks models.

What would settle it

Take a simple binary classifier with known centroid samples x_a, x_b and boundary sample x_b^a. Sample a probe dataset from a different distribution (e.g., random noise or another domain), compute the quantities in Eq. (12) using probes that the model does not score near plus or minus one, and check whether the resulting r_b^a matches the true r_b^a within tolerance. If it does not, the reconstruction claimed in Lemma 1 fails. Empirically, retrieving models with probe datasets deliberately drawn far from decision boundaries should collapse the reported 94.82% top-1 accuracy if the claim is false.

Watch

Extended reading notes

Core claim

Know2Vec treats knowledge as what a model has learned from its training data, and claims this knowledge is encapsulated by the Knowledge Representation Matrix (KRM): for every class pair, the perturbation vector from a class centroid to a decision-boundary sample. Since training centroids are usually unavailable, Lemma 1 asserts that external probe samples close to the centroids' decision values can be used to reconstruct the same perturbation vectors up to small offsets. The KRM is expanded into per-class directed graphs, and an inner-outer bidirectional LSTM encodes these graphs into a model vector h. Query tasks are encoded by averaging class samples and running a bidirectional LSTM over class means, producing vector t. A supervised alignment space, trained with a model-embedding consistency loss and a cosine-margin spatial alignment loss, makes cosine distance between t and h rank the best model; the paper reports superior retrieval accuracy against state-of-the-art baselines on both neural network retrieval and source-free transferability estimation tasks.

Load-bearing premise

The whole scheme rests on Lemma 1's assumption that arbitrary probe datasets contain samples the model scores arbitrarily close to its training-centroid values, so that perturbation vectors computed from probes match those computed from real training centroids.

Editorial extensions

If this is right

  • Model marketplace users can retrieve a fine-tune-ready model by submitting only a small labeled query task, without uploading full data or requiring white-box model access.
  • Model owners can keep parameters and training data private while remaining searchable through a probe-based index.
  • If Lemma 1 holds for arbitrary probes, the proxy generalizes to any model zoo regardless of the models' original training domains.
  • The same vector space supports both retrieval (top-k accuracy) and transferability ranking (Pearson and Spearman) in one framework.
  • The learned proxy can be precomputed offline, making online query-time retrieval nearly instant while large-language-model selection remains slow and weak.

Reading between the lines

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

  • If the probe-substitution proof is unsound, the method may still work empirically because the supervised alignment loss and the model-index classification loss could be absorbing probe-domain noise; this would make the result a learned heuristic rather than a proven black-box characterization.
  • A stronger benchmark would swap probe datasets between retrieval and training time and measure robustness; the paper's own ablation of training versus alternative probes hints that probe choice matters less than claimed, but only on a narrow set.
  • The boundary-sample vectors double as a model signature, so the framework suggests a natural extension to model fingerprinting and intellectual-property protection.
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 / 6 minor

Summary. The paper proposes Know2Vec, a black-box proxy for neural network retrieval. It first constructs a Knowledge Representation Matrix (KRM) for each candidate model from decision-boundary perturbation vectors, using external probe datasets rather than the model's training data. These KRMs are encoded as graphs and embedded via a bidirectional LSTM architecture; query tasks are embedded by averaging per-class features and feeding them through another LSTM. A learned alignment space, trained with a cross-entropy consistency loss and a cosine-margin spatial alignment loss, matches query vectors to model vectors at inference. Experiments on NNR and SF-MTE benchmarks report superior retrieval accuracy and transferability correlation over several baselines, together with ablations on architecture choices, loss functions, and probe datasets. The central theoretical claim is Lemma 1, which purports to prove that perturbation vectors in the KRM can be obtained from external datasets.

Significance. If the claims were established, Know2Vec would be a practically useful, privacy-preserving model retrieval method: it requires only black-box access to models, avoids training-data disclosure, and reports strong results across diverse tasks. The paper contains a substantial experimental effort, comparisons with many baselines, visualizations, and an available code repository. However, the theoretical foundation is not sound as written: the proof of Lemma 1 contains an algebraic inconsistency and uses the Mean Value Theorem in a way that does not establish the required input-space relation. The empirical results are also reported without error bars or significance tests, so the claimed advantages are not yet fully supported. Given that the proof of probe-based KRM replacement is a load-bearing contribution, the manuscript is not ready for publication in its current form, although the empirical approach may be salvageable after substantial revision.

major comments (3)
  1. [Lemma 1, Proof 1, Eq. (12)] The proof of Lemma 1 does not establish that perturbation vectors can be recovered from external probe samples. The Mean Value Theorem only guarantees the existence of a point where the derivative equals a difference quotient; it does not imply that the input offset σ = z_b^a - x_b^a is small, nor does it relate σ to the output differences λ1, λ2, λ3. Moreover, Eq. (12) is algebraically inconsistent with the preceding definitions: from z_b^a = x_b^a + σ and x_a = z_a + σ_a, the identity is r_b^a = z_b^a - z_a - σ - σ_a, not z_b^a - z_a + σ + σ_a. For a network with steep or curved decision boundaries, small output differences can correspond to arbitrarily large input displacements, so the reconstructed r_b^a is not shown to approximate the true KRM vector.
  2. [Lemma 1, Proof 1; 'Using probe datasets instead of training datasets'] The proof assumes the existence of selected probe samples z_a and z_b with δ(w·z_a+b)=1-λ1 and δ(w·z_b+b)=-1+λ2 for arbitrarily small λ1, λ2, and a boundary probe z_b^a with value -λ3. This is an existential condition on the probe dataset: the probes must contain samples arbitrarily close in output space to the training centroids and to the decision boundary. The paper does not show that randomly selected probes satisfy this condition, and in the NNR experiments the probes are drawn from the Know2Vec training set rather than from the target model's training set. Without this assumption, the reconstructed KRM is not guaranteed to represent the model's knowledge, so the central claim that 'it is feasible to obtain model information with randomly selected probes' (Key Contributions) is unsupported.
  3. [Experiments, Tables 1-4] All experimental results are reported as single-point estimates without error bars, confidence intervals, or significance tests. The claimed 1.72% improvement over the suboptimal baseline in Table 1 and the correlation differences in Table 2 may be within the noise of a single run, especially given the lack of repeated training runs with different random seeds and the small number of SF-MTE downstream tasks (four datasets). The paper should report means and standard deviations over multiple seeds, and ideally a paired significance test across tasks, before the superiority claims can be accepted.
minor comments (6)
  1. [Appendix, 'Implementation of External Probe Datasets'] The exact algorithm for generating boundary samples from probe datasets is deferred to reference (Tian et al. 2023) without being reproduced; since boundary sample generation is central to KRM construction, the paper should specify the procedure or include the relevant code excerpt.
  2. [Ablation Study, Tables 3 and 4] The ablation results are presented without variance information, so it is unclear whether differences such as 94.82% versus 93.53% are statistically meaningful.
  3. [Eq. (15) and Eq. (16)] The margin value of 0.4 and the balancing coefficient α=1 are fixed without sensitivity analysis; a brief study of these hyperparameters would strengthen the alignment-loss discussion.
  4. [Appendix, 'Assumption 1'] The appendix acknowledges that ReLU is not differentiable at zero, but the proof of Lemma 2 still relies on differentiability of δ without explaining how the non-differentiable points are handled for actual networks; this should be clarified.
  5. [Appendix, 'Implementation Details of the Kaggle Model Zoo'] The reference to 'Mobile-Net(?)' contains an unresolved citation placeholder and should be corrected.
  6. [Problem Formulation, Eq. (13)] The notation [s_i * I(l_i = k)] mixes sample and indicator notation; the intended element-wise multiplication should be defined explicitly.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the alignment space is trained with the same loss used at inference, but on held-out query tasks; Lemma 1's gap is a proof-soundness issue, not a construction-level reduction.

full rationale

Know2Vec's retrieval pipeline is an empirical supervised meta-learning setup: model knowledge is vectorized from probes, query tasks are vectorized, and a cosine alignment space is trained with cross-entropy and ranking losses. The fact that inference uses the same cosine nearest-neighbor objective as training is standard practice, not circular, because evaluation is on held-out query tasks; the paper states that there was no dataset overlap among model training datasets, Know2Vec training datasets, and query task sets, so the reported retrieval accuracy is not fitted on the test tasks. The load-bearing premise that decision-boundary samples represent model knowledge is imported from external prior work (Tian et al. 2023), and there is no evidence in this manuscript that the authors of Know2Vec are the authors of that cited theorem; even if there were overlap, the cited work is independently published and externally falsifiable. The main weakness is Lemma 1 and the appendix's Lemma 2: the Mean Value Theorem supplies only a derivative identity, and the proof defines sigma as the unknown difference between probe and training boundary samples, so Eq. (12) still contains training-derived offsets and does not by itself establish that arbitrary probes recover the KRM. This is a serious soundness gap, but it is not an equation-level reduction of the claimed result to a fitted parameter or to a self-citation. The appendix also honestly flags the ReLU differentiability simplification. Therefore no load-bearing step is equivalent to its inputs by construction, and the circularity score is 0.

Assumptions & free parameters 5 free parameters · 4 assumptions · 0 invented entities

The method depends on a cited theorem about decision boundary knowledge, a strong differentiability assumption, and the unproven premise that external probe data can substitute for training data. Loss weights, margins, and vector dimensions are chosen by hand.

free parameters (5)
  • alpha = 1
    Balance weight between L_MKC and L_SAL, fixed to 1 without sensitivity analysis.
  • margin = 0.4
    Margin in the cosine spatial alignment loss, chosen by hand with no sensitivity study.
  • knowledge embedding dimension = 256
    Vector length for model and query embeddings; ablation tests 128, 256 and 512, then 256 is selected.
  • LSTM hidden size = 1000
    Hidden embedding size in all LSTM networks, set without justification.
  • query images per class = 5
    NNR query tasks use 5 randomly sampled images per category; ablation ranges from 2 to 8.
assumptions (4)
  • domain assumption Theorem 1 from Tian et al. 2023: knowledge transferred from training data to a model is represented by the KRM of perturbation vectors.
    The paper uses this citation as the grounding for MEXT without proving or verifying it.
  • ad hoc to paper Assumption 1: delta = gA - gB is differentiable near decision boundary samples.
    The appendix explicitly calls this a strong assumption and notes that ReLU is not differentiable at zero; it underpins the Mean Value Theorem step in Lemma 1.
  • ad hoc to paper External probe samples za and zb exist with decision values 1 - lambda1 and -1 + lambda2 arbitrarily close to training centroid values.
    This premise is needed for replacing training samples with external probes, but it is assumed and not established for arbitrary probe datasets.
  • domain assumption The graph set G_Phi, composed of centroids and boundary samples, preserves enough model knowledge for retrieval after LSTM encoding.
    The paper asserts G_Phi is richer than KRM and that bidirectional LSTM reaches all edges, but no formal guarantee ties this to retrieval accuracy.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Know2Vec: A Black-Box Proxy for Neural Network Retrieval." pith.science (2026). https://pith.science/paper/L5LPCFYD

@misc{pith2026241216251,
  author       = {Pith},
  title        = {Pith review of: Know2Vec: A Black-Box Proxy for Neural Network Retrieval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/L5LPCFYD}},
  note         = {Machine review of arXiv:2412.16251}
}
read the original abstract

For general users, training a neural network from scratch is usually challenging and labor-intensive. Fortunately, neural network zoos enable them to find a well-performing model for directly use or fine-tuning it in their local environments. Although current model retrieval solutions attempt to convert neural network models into vectors to avoid complex multiple inference processes required for model selection, it is still difficult to choose a suitable model due to inaccurate vectorization and biased correlation alignment between the query dataset and models. From the perspective of knowledge consistency, i.e., whether the knowledge possessed by the model can meet the needs of query tasks, we propose a model retrieval scheme, named Know2Vec, that acts as a black-box retrieval proxy for model zoo. Know2Vec first accesses to models via a black-box interface in advance, capturing vital decision knowledge from models while ensuring their privacy. Next, it employs an effective encoding technique to transform the knowledge into precise model vectors. Secondly, it maps the user's query task to a knowledge vector by probing the semantic relationships within query samples. Furthermore, the proxy ensures the knowledge-consistency between query vector and model vectors within their alignment space, which is optimized through the supervised learning with diverse loss functions, and finally it can identify the most suitable model for a given task during the inference stage. Extensive experiments show that our Know2Vec achieves superior retrieval accuracy against the state-of-the-art methods in diverse neural network retrieval tasks.

Figures

Figures reproduced from arXiv: 2412.16251 by the authors.

Figure 1
Figure 1. Knowledge-consistency-based black-box proxy [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Model Retrieval Framework. and clarify the requirements of the tasks. We tackle the new issues that incurred from the limited known information. Next, we seek a knowledge-consistent space that acts as a bridge, which connects the two modalities despite their dif￾ferences in structures and semantic parameters, and provid￾ing a way to measure their semantic similarity. Problem Formulation We consider an arbitrary quer… view at source ↗
Figure 3
Figure 3. Decision Boundary Sample. Theorem 1 (Tian et al. 2023) The knowledge transferred from a training dataset to a deep learning model can be represented by the knowledge representation matrix KRM formed by perturbation vectors across different classes. For a k-class classifier, let the centroid sample of category A be denoted as xa, the perturbation vector r k a = x k a − xa from category A to category K is defined as t… view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Visualization description of Pearson correlation on SF-MTE experiments. [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Visual description of model knowledge vectors. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Retrieval accuracy of different number of query [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Visualization Explanation of Three Query Datasets. Figures (e), (f), and (g) are example images of these semantically [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Visualization of Boundary decision samples of different datasets. [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]
Figure 9
Figure 9. Figure 9: Visualization of model knowledge graph sets. From top to bottom are the knowledge graph sets of three models. From [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

51 extracted references · 27 canonical work pages

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 word.in bbl.in capitalize " " * FUNCT...

  3. [3]

    L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al

    Achiam, J.; Adler, S.; Agarwal, S.; Ahmad, L.; Akkaya, I.; Aleman, F. L.; Almeida, D.; Altenschmidt, J.; Altman, S.; Anadkat, S.; et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774

    arXiv 2023

  4. [4]

    Bao, Y.; Li, Y.; Huang, S.; Zhang, L.; Zheng, L.; Zamir, A.; and Guibas, L. J. 2019. An Information-Theoretic Approach to Transferability in Task Transfer Learning. In 2019 IEEE International Conference on Image Processing, ICIP 2019, Taipei, Taiwan, September 22-25, 2019 , 2309--2313. IEEE

    work page 2019

  5. [5]

    Bossard, L.; Guillaumin, M.; and Van Gool, L. 2014. Food-101--mining discriminative components with random forests. In Computer vision--ECCV 2014: 13th European conference, zurich, Switzerland, September 6-12, 2014, proceedings, part VI 13, 446--461. Springer

    2014

  6. [6]

    Cao, X.; Jia, J.; and Gong, N. Z. 2021. IPGuard: Protecting Intellectual Property of Deep Neural Networks via Fingerprinting the Classification Boundary. In Cao, J.; Au, M. H.; Lin, Z.; and Yung, M., eds., ASIA CCS '21: ACM Asia Conference on Computer and Communications Security, Virtual Event, Hong Kong, June 7-11, 2021 , 14--25. ACM

    work page 2021

  7. [7]

    Cheng, G.; Han, J.; and Lu, X. 2017. Remote sensing image scene classification: Benchmark and state of the art. Proceedings of the IEEE, 105(10): 1865--1883

    2017

  8. [8]

    Cimpoi, M.; Maji, S.; Kokkinos, I.; Mohamed, S.; and Vedaldi, A. 2014. Describing textures in the wild. In Proceedings of the IEEE conference on computer vision and pattern recognition, 3606--3613

    2014

Show all 51 references
  1. [9]

    Coates, A.; Ng, A.; and Lee, H. 2011. An analysis of single-layer networks in unsupervised feature learning. In Proceedings of the fourteenth international conference on artificial intelligence and statistics, 215--223. JMLR Workshop and Conference Proceedings

    2011

  2. [10]

    Cohen, I.; Huang, Y.; Chen, J.; Benesty, J.; Benesty, J.; Chen, J.; Huang, Y.; and Cohen, I. 2009. Pearson correlation coefficient. Noise reduction in speech processing, 1--4

    2009

  3. [11]

    Deng, L. 2012. The MNIST Database of Handwritten Digit Images for Machine Learning Research [Best of the Web]. IEEE Signal Processing Magazine, 29(6): 141--142

    2012

  4. [12]

    C.; Bhotika, R.; Soatto, S.; and Perona, P

    Deshpande, A.; Achille, A.; Ravichandran, A.; Li, H.; Zancato, L.; Fowlkes, C. C.; Bhotika, R.; Soatto, S.; and Perona, P. 2021. A linearized framework and a new benchmark for model selection for fine-tuning. CoRR, abs/2102.00084

    2021 arXiv

  5. [13]

    Ding, Y.; Jiang, B.; Yu, A.; Zheng, A.; and Liang, J. 2024. Which Model to Transfer? A Survey on Transferability Estimation. CoRR, abs/2402.15231

    2024 arXiv

  6. [14]

    Fang, C.; Xu, Y.; and Rockmore, D. N. 2013. Unbiased metric learning: On the utilization of multiple datasets and web images for softening bias. In Proceedings of the IEEE International Conference on Computer Vision, 1657--1664

    2013

  7. [15]

    Hauke, J.; and Kossowski, T. 2011. Comparison of values of Pearson's and Spearman's correlation coefficients on the same sets of data. Quaestiones geographicae, 30(2): 87--93

    2011

  8. [16]

    He, K.; Zhang, X.; Ren, S.; and Sun, J. 2016. Deep Residual Learning for Image Recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016 , 770--778. IEEE Computer Society

    2016

  9. [17]

    Helber, P.; Bischke, B.; Dengel, A.; and Borth, D. 2019. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7): 2217--2226

    2019

  10. [18]

    Ho, Y.; and Wookey, S. 2020. The Real-World-Weight Cross-Entropy Loss Function: Modeling the Costs of Mislabeling. IEEE Access, 8: 4806--4813

    2020

  11. [19]

    Ho - Phuoc, T. 2018. CIFAR10 to Compare Visual Recognition Performance between Deep Neural Networks and Humans. CoRR, abs/1811.07270

    2018 arXiv

  12. [20]

    Huang, G.; Liu, Z.; and Weinberger, K. Q. 2016. Densely Connected Convolutional Networks. CoRR, abs/1608.06993

    2016 arXiv

  13. [21]

    Jeong, W.; Lee, H.; Park, G.; Hyung, E.; Baek, J.; and Hwang, S. J. 2021. Task-Adaptive Neural Network Retrieval with Meta-Contrastive Learning. CoRR, abs/2103.01495

    2021 arXiv

  14. [22]

    Khosla, A.; Jayadevaprakash, N.; Yao, B.; and Li, F.-F. 2011. Novel dataset for fine-grained image categorization: Stanford dogs. In Proc. CVPR workshop on fine-grained visual categorization (FGVC), volume 2

    2011

  15. [23]

    Krause, J.; Stark, M.; Deng, J.; and Fei - Fei, L. 2013. 3D Object Representations for Fine-Grained Categorization. In 2013 IEEE International Conference on Computer Vision Workshops, ICCV Workshops 2013, Sydney, Australia, December 1-8, 2013 , 554--561. IEEE Computer Society

    2013

  16. [24]

    J.; and Bottou, L

    LeCun, Y.; Huang, F. J.; and Bottou, L. 2004. Learning methods for generic object recognition with invariance to pose and lighting. In Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2004. CVPR 2004., volume 2, II--104. IEEE

    2004

  17. [25]

    Li, D.; Yang, Y.; Song, Y.-Z.; and Hospedales, T. M. 2017. Deeper, broader and artier domain generalization. In Proceedings of the IEEE international conference on computer vision, 5542--5550

    2017

  18. [26]

    Li, F.-F.; Rob, F.; and Pietro, P. 2004. Learning generative visual models from few training examples: An incremental bayesian approach tested on 101 object categories. In 2004 conference on computer vision and pattern recognition workshop, 178--178. IEEE

    2004

  19. [27]

    Li, Y.; Jia, X.; Sang, R.; Zhu, Y.; Green, B.; Wang, L.; and Gong, B. 2021. Ranking Neural Checkpoints. In IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2021, virtual, June 19-25, 2021 , 2663--2673. Computer Vision Foundation / IEEE

    2021

  20. [28]

    Maji, S.; Rahtu, E.; Kannala, J.; Blaschko, M.; and Vedaldi, A. 2013. Fine-grained visual classification of aircraft. arXiv preprint arXiv:1306.5151

    2013 arXiv

  21. [29]

    Matthey, L.; Higgins, I.; Hassabis, D.; and Lerchner, A. 2017. dsprites: Disentanglement testing sprites dataset

    2017

  22. [30]

    Y.; et al

    Netzer, Y.; Wang, T.; Coates, A.; Bissacco, A.; Wu, B.; Ng, A. Y.; et al. 2011. Reading digits in natural images with unsupervised feature learning. In NIPS workshop on deep learning and unsupervised feature learning, volume 2011, 4. Granada

    2011

  23. [31]

    V.; Hassner, T.; Seeger, M

    Nguyen, C. V.; Hassner, T.; Seeger, M. W.; and Archambeau, C. 2020. LEEP: A New Measure to Evaluate Transferability of Learned Representations. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event , volume 119 of Pr...

    2020

  24. [32]

    Nilsback, M.-E.; and Zisserman, A. 2008. Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & image processing, 722--729. IEEE

    2008

  25. [33]

    Russakovsky, O.; Deng, J.; Su, H.; Krause, J.; Satheesh, S.; Ma, S.; Huang, Z.; Karpathy, A.; Khosla, A.; Bernstein, M.; et al. 2015. Imagenet large scale visual recognition challenge. International journal of computer vision, 115: 211--252

    2015

  26. [34]

    Singla, S.; Singla, S.; and Feizi, S. 2022. Improved deterministic l2 robustness on CIFAR-10 and CIFAR-100 . In The Tenth International Conference on Learning Representations, ICLR 2022, Virtual Event, April 25-29, 2022 . OpenReview.net

    2022

  27. [35]

    Szegedy, C.; Vanhoucke, V.; Ioffe, S.; Shlens, J.; and Wojna, Z. 2016. Rethinking the Inception Architecture for Computer Vision. In 2016 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2016, Las Vegas, NV, USA, June 27-30, 2016 , 2818--2826. IEEE Computer Society

    2016

  28. [36]

    Targ, S.; Almeida, D.; and Lyman, K. 2016. Resnet in resnet: Generalizing residual architectures. arXiv preprint arXiv:1603.08029

    2016 arXiv

  29. [37]

    M.; Hauth, A.; et al

    Team, G.; Anil, R.; Borgeaud, S.; Wu, Y.; Alayrac, J.-B.; Yu, J.; Soricut, R.; Schalkwyk, J.; Dai, A. M.; Hauth, A.; et al. 2023. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805

    2023 arXiv

  30. [38]

    M.; Liu, G.; and Wang, C

    Tian, Z.; Wang, Z.; Abdelmoniem, A. M.; Liu, G.; and Wang, C. 2023. Knowledge Representation of Training Data With Adversarial Examples Supporting Decision Boundary. IEEE Transactions on Information Forensics and Security, 18: 4116--4127

    2023

  31. [39]

    T.; Nguyen, C

    Tran, A. T.; Nguyen, C. V.; and Hassner, T. 2019. Transferability and Hardness of Supervised Classification Tasks. In 2019 IEEE/CVF International Conference on Computer Vision, ICCV 2019, Seoul, Korea (South), October 27 - November 2, 2019 , 1395--1405. IEEE

    2019

  32. [40]

    Van Horn, G.; Branson, S.; Farrell, R.; Haber, S.; Barry, J.; Ipeirotis, P.; Perona, P.; and Belongie, S. 2015. Building a bird recognition app and large scale dataset with citizen scientists: The fine print in fine-grained dataset collection. In Proceedings of the IEEE confer...

    2015

  33. [41]

    Venkateswara, H.; Eusebio, J.; Chakraborty, S.; and Panchanathan, S. 2017. Deep Hashing Network for Unsupervised Domain Adaptation. In 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 5385--5394

    2017

  34. [42]

    Wah, C.; Branson, S.; Welinder, P.; Perona, P.; and Belongie, S. 2011. The caltech-ucsd birds-200-2011 dataset

    2011

  35. [43]

    Xia, G.; Hu, J.; Hu, F.; Shi, B.; Bai, X.; Zhong, Y.; and Zhang, L. 2016. AID: A Benchmark Dataset for Performance Evaluation of Aerial Scene Classification. CoRR, abs/1608.05167

    2016 arXiv

  36. [44]

    A.; Oliva, A.; and Torralba, A

    Xiao, J.; Hays, J.; Ehinger, K. A.; Oliva, A.; and Torralba, A. 2010. SUN database: Large-scale scene recognition from abbey to zoo. In The Twenty-Third IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2010, San Francisco, CA, USA, 13-18 June 2010 , 3485--3492....

    2010

  37. [45]

    You, K.; Liu, Y.; Wang, J.; and Long, M. 2021. LogME: Practical Assessment of Pre-trained Models for Transfer Learning. In Meila, M.; and Zhang, T., eds., Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event , volume 1...

    2021

  38. [46]

    Yu, Y.; Si, X.; Hu, C.; and Zhang, J. 2019. A review of recurrent neural networks: LSTM cells and network architectures. Neural computation, 31(7): 1235--1270

    2019

  39. [47]

    Zhang, Y.; Huang, T.; Ding, Y.; Zhan, D.; and Ye, H. 2023. Model Spider: Learning to Rank Pre-Trained Models Efficiently. In Oh, A.; Naumann, T.; Globerson, A.; Saenko, K.; Hardt, M.; and Levine, S., eds., Advances in Neural Information Processing Systems 36: Annual Conference...

    2023

  40. [48]

    Zhang, Z.; Song, Y.; and Qi, H. 2017. Age progression/regression by conditional adversarial autoencoder. In Proceedings of the IEEE conference on computer vision and pattern recognition, 5810--5818

    2017

  41. [49]

    Zhong, N.; Qian, Z.; and Zhang, X. 2021. Deep Neural Network Retrieval. In Shen, H. T.; Zhuang, Y.; Smith, J. R.; Yang, Y.; C \' e sar, P.; Metze, F.; and Prabhakaran, B., eds., MM '21: ACM Multimedia Conference, Virtual Event, China, October 20 - 24, 2021 , 3455--3463. ACM

    2021

  42. [50]

    Zhou, Z. 2016. Learnware: on the future of machine learning. Frontiers Comput. Sci., 10(4): 589--590

    2016

  43. [51]

    P \' a ndy, M.; Agostinelli, A.; Uijlings, J. R. R.; Ferrari, V.; and Mensink, T. 2022. Transferability Estimation using Bhattacharyya Class Separability. In IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022, New Orleans, LA, USA, June 18-24, 2022 , 916...

    2022

Pith tools

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