Pith. sign in

REVIEW 4 major objections 7 minor 47 references

Interpretable Zero-shot Learning with Infinite Class Concepts

T0 review · 4 major / 7 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read The paper claims that LLM-generated phrase concepts, filtered by an entropy score, can replace finite human-attribute annotations in zero-shot learning while making the class embedding interpretable.

desk verdict A coherent pipeline for LLM-generated concept-based ZSL with consistent but modest gains; the 'image-grounded' claim outruns the evidence. read the letter →

arxiv 2505.03361 v1 pith:RWABOFCC submitted 2025-05-06 cs.CV

Zihan Ye , Shreyank N Gowda , Shiming Chen , Yaochu Jin , Kaizhu Huang , Xiaobo Jin This is my paper · ORCID
classification cs.CV
keywords zero-shotlearningclassconceptslargelanguagemodelshallucinationconceptentropyinterpretabilitysemanticembedding
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 tries to show that zero-shot learning need not depend on expensive, finite human-attribute annotations: a large language model can generate any number of phrase-level visual concepts, and an entropy-based filter can keep only those that are both transferable across classes and discriminative between them. InfZSL builds the class semantic embedding from co-occurrence counts of the surviving concepts, then concatenates it with the human-annotated embedding and plugs into existing embedding or generative zero-shot methods. A sympathetic reader would care because, if true, each dimension of the class embedding becomes a human-readable phrase like 'long neck' rather than an opaque document vector, and the method reports consistent improvements over its host backbones on AWA2, CUB, and SUN without using unseen-class image supervision.

What carries the argument

The load-bearing object is the Concept Entropy Selection and Scoring (CESS) procedure. CESS first clusters LLM-generated concepts using GloVe embeddings and k-means, then for each cluster computes a class-cluster co-occurrence matrix, derives a concept-entropy transferability score and a ktop-th-probability discriminability score, and ranks clusters by the harmonic mean of the two scores. Selected clusters are scored by class-concept co-occurrence counts to form the LLM-concept semantic embedding, which is concatenated with the human-annotated embedding before training the zero-shot classifier.

What would settle it

A concrete failing observation would be a selected concept that is consistently invisible in the class's images, for example 'cunning gaze' for fox showing no stable visual correlate in a ground-truth attribute set. An experiment could take the ten concepts from the paper's user study, annotate their presence in 50 randomly sampled images per class, and check whether the CESS-selected concepts are genuinely more visually present than the eliminated ones; near-chance separation would refute the image-grounded claim.

Watch

Extended reading notes

Core claim

The central claim is that an LLM can supply an effectively infinite set of phrase-level class concepts that, once filtered and scored by the Concept Entropy Selection and Scoring (CESS) procedure, produce a semantic embedding that is at least as accurate as human-annotated attribute embeddings and far more transparent. The paper argues that hallucinated or non-visual phrases like 'cunning behavior' can be detected at the individual-concept level by concept entropy, which measures how evenly a phrase's co-occurrences spread over classes, and that this signal is complementary to a discriminability score based on the ktop-th class probability. Combining the two in a harmonic mean lets the method keep only concepts like 'long neck' that are both shared across classes and informative within them, and the surviving concepts form the class embedding through class-concept co-occurrence counts.

Load-bearing premise

The entire pipeline assumes that how often an LLM attaches a phrase to a class name is a faithful proxy for whether that phrase is visually present in images of the class; if an LLM hallucinates a class-specific but non-visual phrase, it will look discriminative and pass the entropy filter.

Editorial extensions

If this is right

  • InfZSL can be integrated into both embedding and generative zero-shot backbones; the paper reports that it improves its host methods on all three benchmarks, with the largest gains on AWA2.
  • The final class embedding is interpretable: every dimension corresponds to a phrase such as 'small dorsal fin' or 'long vibrissae', and the paper shows attention maps that localize these phrases in images.
  • The transferability-discriminability tradeoff suggests a general recipe for mining visual semantics from LLMs: filter by a harmonic-mean importance score rather than by raw frequency or by a single fidelity score.
  • Because the pipeline needs only class names, LLM prompts, and GloVe embeddings, it could scale to datasets where human attribute annotation is impractical or too costly.

Reading between the lines

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

  • If CESS truly filters hallucinated concepts using only text statistics, the same proxy could plausibly be applied to other vision-language problems such as concept bottleneck models or explainable retrieval, though the paper does not test those settings.
  • A direct extension would be to run CESS on classes that already have visual attribute ground truth, such as CUB, and check whether eliminated concepts are systematically less visually present than selected ones; the paper's own user study covers only four classes and ten concepts.
  • Because the method concatenates LLM and human embeddings, some of the reported gain could come from added embedding dimensions rather than from the filtering itself; a dimension-matched ablation would isolate the contribution of CESS.
  • The 'image-grounded' claim is testable only against pixel-level evidence, so a larger human study or an automatic visual-grounding metric would be the natural next check.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper proposes InfZSL, a zero-shot learning framework that replaces human-annotated attribute vectors with LLM-generated phrase concepts. It generates many concepts per class via a prompt template, clusters them with GloVe embeddings and k-means, computes a class-cluster co-occurrence matrix O, and defines two scores derived from O: a 'concept entropy' transferability score and a ktop-order-probability discriminability score. The harmonic mean of the two ranks the clusters, the top kselect clusters are used to build a class embedding S_m, and S_m is concatenated with human-annotated attributes and fed into an embedding method (I2CFormer) or a generative method (ZeroDiff). Experiments on AWA2, CUB, and SUN report consistent improvements over several concept-based and document-based ZSL baselines, with ablations indicating that each component contributes.

Significance. If the results hold, InfZSL would be a useful step toward reducing annotation cost while retaining interpretability in ZSL. The paper's strengths are a clean three-stage pipeline, comparisons against current concept-based and document-based baselines on three standard benchmarks, component ablations showing the contributions of transferability and discriminability selection, and the use of GloVe rather than CLIP to avoid leaking unseen-class visual information. The improvements over the generative baseline ZeroDiff, however, are small, and the central 'image-grounded' claim currently rests on a single small user study and on the unvalidated assumption that LLM co-occurrence counts proxy visual presence.

major comments (4)
  1. [Section 3.3, Eqs. (1)-(4); Section 3.4] The co-occurrence matrix O is built by counting how often each LLM-generated concept cluster appears in the LLM's concept lists for each class, so both the selection scores and the final embedding S_m are computed from the same LLM text counts and no image information enters the pipeline. The paper's claim that selected concepts are 'image-grounded' is therefore unsupported; the only direct validation, the user study in Appendix A.5, covers 4 classes, 10 concepts, and 29 volunteers, which is too small to bear the weight of the claim. Please add a larger image-based validation (e.g., human ratings of concept presence in images across many classes, or an automated visual check that does not leak unseen-class data) or substantially weaken the 'image-grounded' wording.
  2. [Section 3.3, Eq. (1)] The definition of o_{i,j} is underspecified: the text calls it a 'class-cluster co-occurrence percent' but does not state the exact normalization (per class, per cluster, across the ktime sampled concept lists, after or before deduplication of concepts within a class). Because Eq. (1) applies a softmax to o directly, the chosen normalization changes p(y|q) and hence both r_tran and r_dis, so the reported selection results cannot be reproduced without this information.
  3. [Section 3.3.3] The concept importance r_q is defined as the harmonic mean of a transferability score that is normalized across clusters and a discriminability score that is an order statistic of the softmax probabilities and is not normalized across clusters. These quantities have different scales, and the harmonic mean is introduced without derivation or comparison to alternative fusion functions (e.g., product, weighted sum, min). Add an ablation of the fusion function or a theoretical justification, and report whether the selected concept sets and final accuracies are sensitive to this choice.
  4. [Section 4, Table 1] The gains over the strongest baselines are small in several cases (e.g., InfZSL+ZeroDiff over ZeroDiff is +0.2 H on AWA2 and +0.7 H on CUB), and no error bars, multiple seeds, or significance tests are reported, so the claimed 'significant improvements' are not established. Similarly, the hyperparameters (kpre, kselect, ktop) are chosen empirically per dataset without a validation protocol, and the sensitivity analysis in Fig. 4 covers only AWA2 and CUB, not SUN. Report run-to-run variance and a validation-based selection procedure for hyperparameters.
minor comments (7)
  1. [Section 3.4] The phrase 'resulting our our concept-based semantic embeddings' contains a duplicated 'our'; also correct 'intergrated' in Section 4 and 'promt' in Section 3.2.
  2. [Figure 4] The axis label 'AUSUC' is never defined in the text; please define it or rename the metric.
  3. [Section 4.1] The text says InfZSL achieves 'best performance 76.5%' on AWA2 ZSL, but Table 1 reports 76.6; please correct the inconsistency.
  4. [Eq. (4)] The notation '`p' for the sorted probability sequence is not defined; please define it explicitly.
  5. [Figure 12] The caption refers to 'I2DFormer' in the architecture description, but the model in the figure and text is I2CFormer.
  6. [Section 4.2.1] The statement that 'Inf.'+'Tran.' generally have large improvements overstates the numbers in Table 2; the differences are 0.7-2.7 points and not consistently 'large'.
  7. [Section 1] The claim that the method 'sets a new standard for interpretable ZSL' is too strong given the limited grounding validation.

Circularity Check

1 steps flagged · score 4.0 of 10

CESS selection and scoring both derive from the same LLM co-occurrence counts, so the 'image-grounded' claim is an assumption rather than an independent result.

  1. fitted input called prediction [Section 3.3 (Eqs. 1-4), Section 3.4, and Abstract]
    "we then calculate the class-cluster co-occurrence percent oi,j∈O as the cluster-conditioned class probability ... (1) ... we also count the number of occurrences of selected concept clusters Q' for every classes and use the mean value along classes, resulting our concept-based semantic embeddings Sm. ... Our InfZSL framework ... generates highly interpretable, image-grounded concepts."

    The 'goodness' filter (transferability and discriminability) is computed exclusively from the class-cluster co-occurrence matrix o, which counts how often each LLM-generated concept cluster appears in LLM concept lists. The final Sm is then built by counting co-occurrences of the selected clusters for each class. No image data enters either step. Therefore the claimed 'image-grounded' concepts are not derived from images; they are a re-expression of the same LLM text co-occurrence statistics used to define 'goodness.' A hallucinated phrase that is class-correlated in the LLM output will have moderate entropy and pass the harmonic-mean filter, so the entropy filter cannot distinguish genuine visual attributes from text-correlated hallucinations.

full rationale

The numerical ZSL/GZSL evaluations on AWA2, CUB, and SUN are self-contained against external benchmarks and do not reduce to the method's inputs: the classifier is trained on real images and evaluated on unseen classes, so the reported accuracy gains are independent evidence. However, the paper's interpretability claim ('image-grounded concepts') is undermined by a self-referential selection loop: transferability and discriminability are both defined on LLM class-concept co-occurrences, and the selected semantic embedding is scored from the same counts. This makes the hallucination filter unable to detect class-correlated hallucinations and makes 'image-grounding' an assumption rather than a derived property. This is a partial circularity affecting the central interpretability claim, but not the benchmark results, hence score 4.

Assumptions & free parameters 6 free parameters · 3 assumptions · 0 invented entities

The central contribution relies on LLM-generated concept statistics, and all selection and scoring hyperparameters are chosen by hand per dataset. No external image grounding is used to validate the selected concepts during development; the user study is limited to four classes and ten concepts.

free parameters (6)
  • knum = 100
    Number of visual concepts requested per LLM prompt; set empirically for all datasets.
  • kmax = 3
    Maximum words per generated concept; hand-selected because longer concepts may reduce transferability.
  • ktime = 5 (AWA2/CUB), 1 (SUN)
    Number of repeated LLM generations per class; chosen empirically due to differing class counts.
  • kpre = 200 (AWA2), 500 (CUB), 200 (SUN)
    Number of k-means clusters for concept grouping; tuned per dataset.
  • kselect = 60 (AWA2), 200 (CUB), 100 (SUN)
    Number of concept clusters retained after importance ranking; tuned per dataset.
  • ktop = 3 (AWA2), 10 (CUB), 10 (SUN)
    Order statistic used in the discriminability score in Eq. (4); tuned per dataset.
assumptions (3)
  • domain assumption LLM-generated concept co-occurrence across class names is a faithful proxy for visual co-occurrence in images, so concepts that are discriminative in class labels are also visually discriminative.
    Section 3.3 uses p(y|q) computed from LLM-generated co-occurrences to select concepts and build S_m; no image evidence is used in the selection.
  • domain assumption GloVe word-vector similarity is sufficient to cluster phrase-level visual concepts into semantically meaningful groups.
    Section 3.3 clusters concepts by mean GloVe embeddings; the paper deliberately avoids CLIP to prevent unseen-class leakage, but GloVe may not capture visual similarity of phrases.
  • ad hoc to paper The harmonic mean of normalized concept entropy and the ktop-th order probability is a valid importance score for balancing transferability and discriminability.
    The formula r_q = 2 * r_tran * r_dis / (r_tran + r_dis) is introduced without derivation; the sensitivity analysis only varies the hyperparameters, not the functional form.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Interpretable Zero-shot Learning with Infinite Class Concepts." pith.science (2026). https://pith.science/paper/RWABOFCC

@misc{pith2026250503361,
  author       = {Pith},
  title        = {Pith review of: Interpretable Zero-shot Learning with Infinite Class Concepts},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/RWABOFCC}},
  note         = {Machine review of arXiv:2505.03361}
}
read the original abstract

Zero-shot learning (ZSL) aims to recognize unseen classes by aligning images with intermediate class semantics, like human-annotated concepts or class definitions. An emerging alternative leverages Large-scale Language Models (LLMs) to automatically generate class documents. However, these methods often face challenges with transparency in the classification process and may suffer from the notorious hallucination problem in LLMs, resulting in non-visual class semantics. This paper redefines class semantics in ZSL with a focus on transferability and discriminability, introducing a novel framework called Zero-shot Learning with Infinite Class Concepts (InfZSL). Our approach leverages the powerful capabilities of LLMs to dynamically generate an unlimited array of phrase-level class concepts. To address the hallucination challenge, we introduce an entropy-based scoring process that incorporates a ``goodness" concept selection mechanism, ensuring that only the most transferable and discriminative concepts are selected. Our InfZSL framework not only demonstrates significant improvements on three popular benchmark datasets but also generates highly interpretable, image-grounded concepts. Code will be released upon acceptance.

Figures

Figures reproduced from arXiv: 2505.03361 by the authors.

Figure 1
Figure 1. Motivation Illustration. (a) LLM-based document ZSL methods encounter two main issues: [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. An illustration of our InfZSL. It consists of three steps. (a) Concept generation: we use [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. The heatmap visualization of our LLM-concept-based semantic embeddings. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (8 more)
Figure 5
Figure 5. Figure 5: Attention visualizations of our LLM￾generated concepts for two unseen classes. 4.3 Qualitative Results LLM-generated Concepts We provide the results about generated concepts and corresponding document-version in Appendix A.2. The results exhibit that, even we explicitl…
Figure 6
Figure 6. Figure 6: The examples of LLM-generated concepts on AWA2. [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The examples of LLM-generated concepts on CUB. [PITH_FULL_IMAGE:figures/full_fig_p013_7.png]
Figure 8
Figure 8. Figure 8: The examples of LLM-generated documents on AWA2. [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: The effect of varying ktop. 14 [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: Attention visualizations of human-annotated concepts for four unseen classes. [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: User study to our (a) selected and (b) eliminated concepts. [PITH_FULL_IMAGE:figures/full_fig_p015_11.png]
Figure 12
Figure 12. Figure 12: The detailed architecture of our I2DFormer. [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

47 extracted references · 28 canonical work pages

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774, 2023. 9

    arXiv 2023

  2. [2]

    k-means++: The advantages of careful seeding

    David Arthur and Sergei Vassilvitskii. k-means++: The advantages of careful seeding. Technical report, Stanford, 2006

    2006

  3. [3]

    Can we open the black box of ai? Nature News, 538(7623):20, 2016

    Davide Castelvecchi. Can we open the black box of ai? Nature News, 538(7623):20, 2016

    work page 2016

  4. [4]

    Transzero++: Cross attribute-guided transformer for zero-shot learning

    Shiming Chen, Ziming Hong, Wenjin Hou, Guo-Sen Xie, Yibing Song, Jian Zhao, Xinge You, Shuicheng Yan, and Ling Shao. Transzero++: Cross attribute-guided transformer for zero-shot learning. IEEE transactions on pattern analysis and machine intelligence, 2022

    work page 2022

  5. [5]

    Transzero: Attribute-guided transformer for zero-shot learning

    Shiming Chen, Ziming Hong, Yang Liu, Guo-sen Xie, Baigui Sun, Hao Li, Qinmu Peng, Ke Lu, and Xinge You. Transzero: Attribute-guided transformer for zero-shot learning. In AAAI, 2022

    work page 2022

  6. [6]

    Evolving semantic prototype improves generative zero-shot learning

    Shiming Chen, Wenjin Hou, Ziming Hong, Xiaohan Ding, Yibing Song, Xinge You, Tongliang Liu, and Kun Zhang. Evolving semantic prototype improves generative zero-shot learning. In Proceedings of the 40th International Conference on Machine Learning , volume 202 of Proceedings of Machine Learning Research, pages 4611–4622. PMLR, 23–29 Jul 2023

    work page 2023

  7. [7]

    Progressive semantic- guided vision transformer for zero-shot learning

    Shiming Chen, Wenjin Hou, Salman Khan, and Fahad Shahbaz Khan. Progressive semantic- guided vision transformer for zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23964–23974, 2024

    work page 2024

  8. [8]

    Hsva: Hierarchical semantic-visual adaptation for zero-shot learning

    Shiming Chen, Guosen Xie, Yang Liu, Qinmu Peng, Baigui Sun, Hao Li, Xinge You, and Ling Shao. Hsva: Hierarchical semantic-visual adaptation for zero-shot learning. Advances in Neural Information Processing Systems, 34, 2021

    work page 2021

Show all 47 references
  1. [9]

    Duet: Cross-modal semantic grounding for contrastive zero-shot learning

    Zhuo Chen, Yufeng Huang, Jiaoyan Chen, Yuxia Geng, Wen Zhang, Yin Fang, Jeff Z Pan, and Huajun Chen. Duet: Cross-modal semantic grounding for contrastive zero-shot learning. In Proceedings of the AAAI conference on artificial intelligence, volume 37, pages 405–413, 2023

    2023

  2. [10]

    Imagenet: A large- scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large- scale hierarchical image database. In 2009 IEEE conference on computer vision and pattern recognition, pages 248–255. Ieee, 2009

    2009

  3. [11]

    Detecting hallucinations in large language models using semantic entropy

    Sebastian Farquhar, Jannik Kossen, Lorenz Kuhn, and Yarin Gal. Detecting hallucinations in large language models using semantic entropy. Nature, 630(8017):625–630, 2024

    2024

  4. [12]

    Controlled hallucinations: Learning to generate faithfully from noisy data

    Katja Filippova. Controlled hallucinations: Learning to generate faithfully from noisy data. arXiv preprint arXiv:2010.05873, 2020

    2010 arXiv

  5. [13]

    Visual-augmented dynamic semantic prototype for generative zero-shot learning

    Wenjin Hou, Shiming Chen, Shuhuang Chen, Ziming Hong, Yan Wang, Xuetao Feng, Salman Khan, Fahad Shahbaz Khan, and Xinge You. Visual-augmented dynamic semantic prototype for generative zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern ...

    2024

  6. [14]

    Survey of hallucination in natural language generation

    Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12):1–38, 2023

    2023

  7. [15]

    Evaluating open-domain question answering in the era of large language models

    Ehsan Kamalloo, Nouha Dziri, Charles Clarke, and Davood Rafiei. Evaluating open-domain question answering in the era of large language models. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages 5591–5606, 2023

    2023

  8. [16]

    Goal- oriented gaze estimation for zero-shot learning

    Yang Liu, Lei Zhou, Xiao Bai, Yifei Huang, Lin Gu, Jun Zhou, and Tatsuya Harada. Goal- oriented gaze estimation for zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 3794–3803, 2021

    2021

  9. [17]

    On faithfulness and factuality in abstractive summarization

    Joshua Maynez, Shashi Narayan, Bernd Bohnet, and Ryan McDonald. On faithfulness and factuality in abstractive summarization. arXiv preprint arXiv:2005.00661, 2020

    2005 arXiv

  10. [18]

    Distributed repre- sentations of words and phrases and their compositionality

    Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. Distributed repre- sentations of words and phrases and their compositionality. Advances in neural information processing systems, 26, 2013. 10

    2013

  11. [19]

    Domain-aware visual bias eliminating for generalized zero-shot learning

    Shaobo Min, Hantao Yao, Hongtao Xie, Chaoqun Wang, Zheng-Jun Zha, and Yongdong Zhang. Domain-aware visual bias eliminating for generalized zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12664–12673, 2020

    2020

  12. [20]

    I2mvformer: Large language model generated multi-view document supervision for zero-shot image classification

    Muhammad Ferjad Naeem, Muhammad Gul Zain Ali Khan, Yongqin Xian, Muhammad Zeshan Afzal, Didier Stricker, Luc Van Gool, and Federico Tombari. I2mvformer: Large language model generated multi-view document supervision for zero-shot image classification. In Proceedings of the IEE...

    2023

  13. [21]

    I2dformer: Learning image to document attention for zero-shot image classification

    Muhammad Ferjad Naeem, Yongqin Xian, Luc V Gool, and Federico Tombari. I2dformer: Learning image to document attention for zero-shot image classification. Advances in Neural Information Processing Systems, 35:12283–12294, 2022

    2022

  14. [22]

    Sun attribute database: Discovering, annotating, and recognizing scene attributes

    Genevieve Patterson and James Hays. Sun attribute database: Discovering, annotating, and recognizing scene attributes. In 2012 IEEE Conference on Computer Vision and Pattern Recognition, pages 2751–2758. IEEE, 2012

    2012

  15. [23]

    The human imagination: the cognitive neuroscience of visual mental imagery

    Joel Pearson. The human imagination: the cognitive neuroscience of visual mental imagery. Nature reviews neuroscience, 20(10):624–634, 2019

    2019

  16. [24]

    Glove: Global vectors for word representation

    Jeffrey Pennington, Richard Socher, and Christopher D Manning. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532–1543, 2014

    2014

  17. [25]

    Less is more: zero-shot learning from online textual documents with noise suppression

    Ruizhi Qiao, Lingqiao Liu, Chunhua Shen, and Anton Van Den Hengel. Less is more: zero-shot learning from online textual documents with noise suppression. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 2249–2257, 2016

    2016

  18. [26]

    Visual-semantic decomposition and partial alignment for document-based zero-shot learning

    Xiangyan Qu, Jing Yu, Keke Gai, Jiamin Zhuang, Yuanmin Tang, Gang Xiong, Gaopeng Gou, and Qi Wu. Visual-semantic decomposition and partial alignment for document-based zero-shot learning. In ACM Multimedia 2024

    2024

  19. [27]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International conference on machine learning, pa...

    2021

  20. [28]

    Generating descriptions with grounded and co-referenced people

    Anna Rohrbach, Marcus Rohrbach, Siyu Tang, Seong Joon Oh, and Bernt Schiele. Generating descriptions with grounded and co-referenced people. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 4979–4989, 2017

    2017

  21. [29]

    Term-weighting approaches in automatic text retrieval

    Gerard Salton and Christopher Buckley. Term-weighting approaches in automatic text retrieval. Information processing & management, 24(5):513–523, 1988

    1988

  22. [30]

    Reinforcement learning from human feedback: Progress and challenges

    John Schulman. Reinforcement learning from human feedback: Progress and challenges. In Berkeley EECS Colloquium. YouTube www. youtube. com/watch, 2023

    2023

  23. [31]

    Incremental residual concept bottleneck models

    Chenming Shang, Shiji Zhou, Hengyuan Zhang, Xinzhe Ni, Yujiu Yang, and Yuwang Wang. Incremental residual concept bottleneck models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 11030–11040, 2024

    2024

  24. [32]

    Selective zero-shot classification with augmented attributes

    Jie Song, Chengchao Shen, Jie Lei, An-Xiang Zeng, Kairi Ou, Dacheng Tao, and Mingli Song. Selective zero-shot classification with augmented attributes. In Proceedings of the European Conference on Computer Vision (ECCV), pages 468–483, 2018

    2018

  25. [33]

    Gemini: a family of highly capable multimodal models

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean-Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023

    2023 arXiv

  26. [34]

    The caltech-ucsd birds-200-2011 dataset

    Catherine Wah, Steve Branson, Peter Welinder, Pietro Perona, and Serge Belongie. The caltech-ucsd birds-200-2011 dataset. 2011. 11

    2011

  27. [35]

    Chain-of-thought prompting elicits reasoning in large language models

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837, 2022

    2022

  28. [36]

    Zero-shot learning—a comprehensive evaluation of the good, the bad and the ugly

    Yongqin Xian, Christoph H Lampert, Bernt Schiele, and Zeynep Akata. Zero-shot learning—a comprehensive evaluation of the good, the bad and the ugly. IEEE transactions on pattern analysis and machine intelligence, 41(9):2251–2265, 2018

    2018

  29. [37]

    Feature generating networks for zero-shot learning

    Yongqin Xian, Tobias Lorenz, Bernt Schiele, and Zeynep Akata. Feature generating networks for zero-shot learning. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5542–5551, 2018

    2018

  30. [38]

    On hallucination and predictive uncertainty in conditional language generation

    Yijun Xiao and William Yang Wang. On hallucination and predictive uncertainty in conditional language generation. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, 2021

    2021

  31. [39]

    Attribute prototype network for zero-shot learning

    Wenjia Xu, Yongqin Xian, Jiuniu Wang, Bernt Schiele, and Zeynep Akata. Attribute prototype network for zero-shot learning. In NeurIPS, 2020

    2020

  32. [40]

    Vgse: Visually- grounded semantic embeddings for zero-shot learning

    Wenjia Xu, Yongqin Xian, Jiuniu Wang, Bernt Schiele, and Zeynep Akata. Vgse: Visually- grounded semantic embeddings for zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9316–9325, 2022

    2022

  33. [41]

    Language in a bottle: Language model guided concept bottlenecks for interpretable image classification

    Yue Yang, Artemis Panagopoulou, Shenghao Zhou, Daniel Jin, Chris Callison-Burch, and Mark Yatskar. Language in a bottle: Language model guided concept bottlenecks for interpretable image classification. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern R...

    2023

  34. [42]

    Exploring data efficiency in zero-shot learning with diffusion models

    Zihan Ye, Shreyank N Gowda, Xiaobo Jin, Xiaowei Huang, Haotian Xu, Yaochu Jin, and Kaizhu Huang. Exploring data efficiency in zero-shot learning with diffusion models. arXiv preprint arXiv:2406.02929, 2024

    2024 arXiv

  35. [43]

    Rebalanced zero-shot learning

    Zihan Ye, Guanyu Yang, Xiaobo Jin, Youfa Liu, and Kaizhu Huang. Rebalanced zero-shot learning. IEEE Transactions on Image Processing, 2023

    2023

  36. [44]

    Designing category-level attributes for discriminative visual recognition

    Felix X Yu, Liangliang Cao, Rogerio S Feris, John R Smith, and Shih-Fu Chang. Designing category-level attributes for discriminative visual recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 771–778, 2013

    2013

  37. [45]

    Benchmarking large language models for news summarization

    Tianyi Zhang, Faisal Ladhak, Esin Durmus, Percy Liang, Kathleen McKeown, and Tatsunori B Hashimoto. Benchmarking large language models for news summarization. Transactions of the Association for Computational Linguistics, 12:39–57, 2024

    2024

  38. [46]

    Grounded video description

    Luowei Zhou, Yannis Kalantidis, Xinlei Chen, Jason J Corso, and Marcus Rohrbach. Grounded video description. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6578–6587, 2019. A Additional Experiments A.1 Additional Ablation Study A.2 ...

    2019

  39. [47]

    Inference

    Our I2CFormer is optimized by Adam optimizer with a learning rate of 0.0005, momentum of 0.9, and weight decay of 0.0001. Inference. During training, the model merely learns about the knowledge of seen categories, whereas both seen and unseen categories are available at infere...

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.