Pith. sign in

REVIEW 4 major objections 6 minor 1 cited by

Discriminative Image Generation with Diffusion Models for Zero-Shot Learning

T0 review · 4 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 A diffusion-based pipeline that generates discriminative images for unseen classes can train a zero-shot classifier, surpassing prior nonhuman-annotated semantic prototype methods by a large margin on four benchmarks.

desk verdict Strong new idea for image-based ZSL, but missing vanilla-SD control and contamination audit leave the headline claim unproven. read the letter →

arxiv 2412.17219 v2 pith:426FTNFN submitted 2024-12-23 cs.CV

Dingjie Fu , Wenjin Hou , Shiming Chen , Shuhuang Chen , Xinge You , Salman Khan , Fahad Shahbaz Khan This is my paper · ORCID
classification cs.CV
keywords zero-shotlearninggenerativetext-to-imagediffusionmodelsdiscriminativeclasstokenssemanticprototypesgeneralized
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

Zero-shot learning normally synthesizes feature vectors for unseen classes from human-annotated attributes. The paper claims this step can be replaced by generating actual images with a text-to-image diffusion model, using only class names as semantic information. To make the images useful for classification, the paper learns a discriminative class token for each unseen class under the guidance of a category discrimination model trained on seen classes. The paper reports that a classifier trained on these generated images plus real seen images sets new state-of-the-art results for nonhuman-annotated semantic prototypes, with an average Top-1 improvement of 24.2% over prior methods, and matches or exceeds methods that use human-annotated attributes on three of four benchmarks. If correct, the result removes the main scalability bottleneck of generative zero-shot learning: no expert attribute annotation is needed for new classes.

What carries the argument

The carrying object is the discriminative class token: a single learned embedding vector added to the tokenizer vocabulary and placed into the diffusion prompt between 'A photo of' and the class name. The token is optimized so that images generated by Stable Diffusion receive high cosine similarity under the category discrimination model to the correct unseen class's CLIP text prototype and low similarity to other unseen classes; an early-stopping threshold on classification accuracy prevents over-training. The category discrimination model provides the only learning signal from seen data, so the whole pipeline needs no human-annotated attributes. The same backbone features are reused when training the final classifier.

What would settle it

Run the exact DIG-ZSL protocol on a set of newly created classes that provably did not exist when Stable Diffusion and CLIP were trained, verify their absence by retrieval-style overlap checks, and measure Top-1 accuracy; if accuracy collapses toward chance level on such held-out classes while staying high on AWA2, CUB, FLO, and SUN, the benchmark result would be explained by pretraining leakage rather than by the proposed mechanism.

Watch

Extended reading notes

Core claim

The central claim is that zero-shot recognition can be driven by image generation: instead of learning a conditional generator of visual features, one can prompt a frozen text-to-image diffusion model with a learned discriminative class token to synthesize photorealistic training images for each unseen class. The token is inserted into the prompt 'A photo of S* [name]' and optimized under the guidance of a category discrimination model, a frozen vision backbone with a trainable projection trained only on seen classes, using a cross-entropy loss on cosine similarities between the generated image's feature and the CLIP text prototypes of the unseen classes. The generated images are then combined with real seen data to train a standard classifier for both conventional and generalized zero-shot settings. The paper demonstrates the claim on AWA2, CUB, FLO, and SUN, reporting substantial improvements over nonhuman-annotated semantic prototype baselines and accuracy competitive with human-annotated attribute methods; it also shows that fine-tuning the category discrimination model on seen data raises accuracy further and that generated images have lower FID than plain Stable Diffusion outputs.

Load-bearing premise

The result depends on Stable Diffusion and CLIP not already having seen images of the benchmark 'unseen' classes during pretraining; if those classes appear in the training data, the generated images and the CLIP features carry class information, and the reported accuracy measures memorization rather than zero-shot transfer.

Editorial extensions

If this is right

  • If the central claim holds, generative zero-shot learning no longer needs human-annotated attributes: class names alone suffice, so new categories can be added without expert annotation cost.
  • Generated images provide interpretable evidence of what the model believes an unseen class looks like, which feature-synthesis methods do not offer.
  • The same pipeline applies to both conventional and generalized zero-shot settings; on the reported benchmarks, it surpasses previous nonhuman-annotated semantic prototype methods and remains competitive with human-annotated ones.
  • The number of generated images per class is a controllable resource; accuracy rises with more samples and saturates around 100 per class, so users can trade generation cost against accuracy.
  • Improving the category discrimination model, for example by fine-tuning on seen classes, is a direct lever that further boosts the final zero-shot accuracy by 1.2 to 5.1 percentage points on the tested datasets.

Reading between the lines

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

  • Inference: Because the method's benchmark numbers could be inflated if Stable Diffusion or CLIP encountered the unseen classes during web-scale pretraining, the zero-shot claim would be strengthened by repeating the protocol on genuinely novel classes invented after the training cutoff.
  • Inference: The token-learning step is decoupled from the classifier, so the same discriminative class tokens could be reused to generate images at any resolution or count later, or to feed other downstream tasks such as retrieval and captioning.
  • Inference: The category discrimination model is modular; replacing the vision backbone and projection with any model that can score generated images against unseen-class semantic prototypes, trained only on seen data, would likely yield similar behavior.
  • Inference: A few-shot extension is natural: a handful of real images from a new class could refine its token, potentially closing the remaining gap on fine-grained datasets like CUB where the paper reports lower accuracy.
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 / 6 minor

Summary. This paper proposes DIG-ZSL, a zero-shot learning method that synthesizes images for unseen classes with Stable Diffusion. In the first stage, a Category Discrimination Model (CDM), composed of a frozen CLIP visual encoder and a trainable MLP, is trained on seen classes to align visual features with CLIP text embeddings. In the second stage, for each unseen class a discriminative class token (DCT) is optimized so that images generated with the prompt "A photo of S* [name]" maximize the CDM-based cosine similarity to that class's CLIP prototype. In the third stage, a classifier is trained on real seen images plus generated unseen images and evaluated under both conventional and generalized ZSL on AWA2, CUB, FLO, and SUN. The paper reports large improvements over nonhuman-annotated semantic prototype baselines and competitive results against human-annotated attribute methods.

Significance. If the results reflect genuine zero-shot transfer, the paper makes a useful contribution: it demonstrates an interpretable image-generation route to ZSL, removes the need for human-annotated attributes, and provides extensive experiments, visualizations, and hyperparameter analyses. The CDM-guided DCT idea is simple and plausible. However, the significance is currently conditional on two unresolved issues: possible contamination of the web-scale pretrained SD/CLIP models with the benchmark unseen classes, and the absence of a vanilla-SD classification control. The paper also omits details of the DCT optimization and of the validation protocol. Until these are addressed, the headline accuracies cannot be interpreted as evidence of zero-shot generalization. No code is released, so reproducibility rests entirely on the written description.

major comments (4)
  1. [§4.1, Implementation Details, with §3.2–§3.4]
  2. [§4.3, Table 4]
  3. [§3.3 and Appendix A]
  4. [§4.1 and §4.5]
minor comments (6)
  1. [§3.1] The heading "Stable Difusion" contains a typo; it should read "Stable Diffusion."
  2. [§1, §4.1, Table 3] The dataset name is inconsistent: "AW A2" appears in the text while "AWA2" appears in tables; please use one canonical form.
  3. [Appendix A] In the "Supervised Classifier Training" paragraph, the text says the default setting is the same as CDM "as shown in Tab. 7," but Tab. 7 reports DCT learning settings; this cross-reference appears to be wrong.
  4. [§4.3, Table 4] Table 4 reports a single FID value per method without seeds or confidence intervals; given the small margins, the comparison would be more informative with multiple runs or error bars.
  5. [§4.2, Table 2] The phrase "as indicated in gray font" cannot be verified in the text version of Table 2; please use a dedicated column or note to identify the large vision-language baselines.
  6. [§4.2, Table 3] Table 3 omits FLO without explanation; please either include FLO results or note why it is excluded.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the pipeline is self-contained, with unseen test labels entering only at final evaluation.

full rationale

DIG-ZSL's derivation chain is not circular. The CDM is trained exclusively on seen-class images and CLIP text prototypes (Section 3.2), the DCT is optimized for each unseen class using only generated images and the CDM's cosine-similarity score against unseen-class prototypes (Eqs. 4-5, Section 3.3), and the final ZSL classifier is trained on generated unseen images plus real seen images (Section 3.4). Unseen test images and labels are never used in the CDM, DCT, or classifier training stages; they appear only in the final evaluation. The paper's self-citations to the authors' prior ZSL works (e.g., [5, 7, 20, 55]) are background and baseline references, and none is invoked as a load-bearing uniqueness theorem or as justification for the central framework. The main external-validity concern, that Stable Diffusion and CLIP may have been pretrained on images from the benchmark unseen classes, is a potential contamination/memorization risk rather than an internal derivational circularity: no equation reduces to another by construction, and no fitted test statistic is renamed as a prediction. Therefore the reported results, even if contestable on contamination grounds, do not arise from a circular argument within the paper itself.

Assumptions & free parameters 4 free parameters · 5 assumptions · 1 invented entities

The central claim rests on several domain assumptions about CLIP alignment, the transferability of the CDM, and the steerability of Stable Diffusion. The most fragile are the unstated differentiability of the generation step and the absence of a contamination check for the pretrained models. The per-dataset hyperparameters gamma and lambda are hand-tuned and directly affect the reported accuracy.

free parameters (4)
  • Discriminative class token embedding e* = learned vector, dimension equals CLIP text token dimension
    Optimized per unseen class via cross-entropy loss (Eq. 5) to align generated images with the class prototype. It is the central learned entity of the method.
  • Early-stopping threshold gamma = 0.4 (CUB, FLO), 0.6 (SUN, AWA2)
    Hand-set per dataset. Determines when DCT optimization stops based on CDM accuracy on generated images. Directly influences the quality of the generated training set.
  • Calibration coefficient lambda = 0.95 (CUB), 0.6 (SUN), 0.9 (FLO), 0.8 (AWA2)
    Hand-set per dataset. Used in GZSL inference (Eq. 6) to rebalance seen and unseen class logits. Affects the reported harmonic mean.
  • Number of generated images per unseen class Ngen = 100
    Set to 100 for all datasets. The hyperparameter analysis in Fig. 5 shows accuracy increases with Ngen, so this choice affects the final numbers.
assumptions (5)
  • domain assumption CLIP text embeddings of class names are valid semantic prototypes for unseen classes.
    The method uses 'A photo of a [name]' with the CLIP text encoder to obtain prototypes, assuming they align with visual features in a shared space.
  • domain assumption A category discrimination model trained only on seen classes provides useful guidance for unseen classes via cosine similarity to unseen prototypes.
    Sections 3.2 and 3.3: the CDM's projected features are compared to unseen class prototypes to compute the DCT loss. This assumes the semantic space generalizes beyond seen classes.
  • domain assumption Stable Diffusion can generate images that, when steered by a token embedding, display discriminative attributes of unseen classes.
    Section 3.3: the generated images serve as training data for the ZSL classifier. Without this, the pipeline cannot work.
  • ad hoc to paper Gradients can be obtained for the token embedding through the stochastic image sampling process.
    Eq. 5 requires optimizing e* through the generated image exi. The paper does not describe a differentiable path or a surrogate gradient, so this is a load-bearing unstated assumption.
  • domain assumption The pretrained models (CLIP and Stable Diffusion) were not trained on the benchmark datasets' unseen classes.
    Section 4.1 uses Stable Diffusion v2.1 and CLIP ViT-B/16 without discussing possible overlap with AWA2, CUB, FLO, or SUN. If the classes were in the pretraining data, the evaluation is not a valid zero-shot test.
invented entities (1)
  • Discriminative class token (DCT) S*
    purpose: A learned token embedding appended to the prompt to make generated images class-discriminative.
    A new token in the prompt vocabulary. Its only validation is within this paper's framework; there is no external falsifiable handle such as a predicted measurable quantity.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Discriminative Image Generation with Diffusion Models for Zero-Shot Learning." pith.science (2026). https://pith.science/paper/426FTNFN

@misc{pith2026241217219,
  author       = {Pith},
  title        = {Pith review of: Discriminative Image Generation with Diffusion Models for Zero-Shot Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/426FTNFN}},
  note         = {Machine review of arXiv:2412.17219}
}
read the original abstract

Generative Zero-Shot Learning (ZSL) methods synthesize class-related features based on predefined class semantic prototypes, showcasing superior performance. However, this feature generation paradigm falls short of providing interpretable insights. In addition, existing approaches rely on semantic prototypes annotated by human experts, which exhibit a significant limitation in their scalability to generalized scenes. To overcome these deficiencies, a natural solution is to generate images for unseen classes using text prompts. To this end, We present DIG-ZSL, a novel Discriminative Image Generation framework for Zero-Shot Learning. Specifically, to ensure the generation of discriminative images for training an effective ZSL classifier, we learn a discriminative class token (DCT) for each unseen class under the guidance of a pre-trained category discrimination model (CDM). Harnessing DCTs, we can generate diverse and high-quality images, which serve as informative unseen samples for ZSL tasks. In this paper, the extensive experiments and visualizations on four datasets show that our DIG-ZSL: (1) generates diverse and high-quality images, (2) outperforms previous state-of-the-art nonhuman-annotated semantic prototype-based methods by a large margin, and (3) achieves comparable or better performance than baselines that leverage human-annotated semantic prototypes. The codes will be made available upon acceptance of the paper.

Figures

Figures reproduced from arXiv: 2412.17219 by the authors.

Figure 1
Figure 1. Schematic of the ZSL method. (a) Previous works ( [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Illustration of our proposed DIG-ZSL framework. We first learn a category discrimination model (CDM) with the training data from seen classes and semantic prototypes derived from class names. Then, we initialize a token S∗, and iteratively modifying the embedding of this input token (denoted as e∗) under the guidance of CDM, using a cross-entropy loss. Equipped with the optimized token for each unseen category, we i… view at source ↗
Figure 3
Figure 3. t-SNE visualization of the DCT embeddings. A dot indi [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (8 more)
Figure 4
Figure 4. Figure 4: Visualization of generated images at the identical step for both the plain Stable Diffusion (SD) and DIG-ZSL. Our DIG-ZSL demonstrates the capability to enhance text-to-image alignment and significantly improve the portrayal of details. We use CUB as an example, with t…
Figure 5
Figure 5. Figure 5: Hyper-parameter sensitivity. Take CUB dataset as an [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: t-SNE visualization of the DCT embeddings for SUN [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]
Figure 7
Figure 7. Figure 7: Visualization of different values of the threshold [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: More visualizations on AWA2 dataset. 2 [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]
Figure 9
Figure 9. Figure 9: More visualizations on CUB dataset. 3 [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: More visualizations on SUN dataset. 4 [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: More visualizations on FLO dataset. 5 [PITH_FULL_IMAGE:figures/full_fig_p016_11.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. GenPrior: Unleashing Text-to-Motion Generative Priors for Zero-Shot Skeleton-based Action Recognition

    cs.CV 2026-08 conditional novelty 7.0 of 10

    Using text-to-motion generation to inject kinematic structure into text prototypes improves zero-shot skeleton action recognition, though most of the reported gain comes from test-time prototype self-refinement.

Reference graph

Works this paper leans on

62 extracted references · 47 canonical work pages · cited by 1 Pith paper

  1. [1]

    Synthetic data from diffusion models improves imagenet classification

    Shekoofeh Azizi, Simon Kornblith, Chitwan Saharia, Mo- hammad Norouzi, and David J Fleet. Synthetic data from diffusion models improves imagenet classification. arXiv preprint arXiv:2304.08466, 2023. 2

    arXiv 2023

  2. [2]

    Long- former: The long-document transformer

    Iz Beltagy, Matthew E Peters, and Arman Cohan. Long- former: The long-document transformer. arXiv preprint arXiv:2004.05150, 2020. 2, 5

    arXiv 2004

  3. [3]

    iedit: Lo- calised text-guided image editing with weak supervision

    Rumeysa Bodur, Erhan Gundogdu, Binod Bhattarai, Tae- Kyun Kim, Michael Donoser, and Loris Bazzani. iedit: Lo- calised text-guided image editing with weak supervision. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 7426–7435, 2024. 2

    work page 2024

  4. [4]

    Muse: Text-to-image generation via masked generative transform- ers

    Huiwen Chang, Han Zhang, Jarred Barber, AJ Maschinot, Jose Lezama, Lu Jiang, Ming-Hsuan Yang, Kevin Mur- phy, William T Freeman, Michael Rubinstein, et al. Muse: Text-to-image generation via masked generative transform- ers. arXiv preprint arXiv:2301.00704, 2023. 3

    arXiv 2023

  5. [5]

    Free: Feature re- finement for generalized zero-shot learning

    Shiming Chen, Wenjie Wang, Beihao Xia, Qinmu Peng, Xinge You, Feng Zheng, and Ling Shao. Free: Feature re- finement for generalized zero-shot learning. In Proceedings of the IEEE/CVF international conference on computer vi- sion, pages 122–131, 2021. 2

    work page 2021

  6. [6]

    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, 34:16622–16634,

    work page

  7. [7]

    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 International Conference on Machine Learning, pages 4611–4622. PMLR, 2023. 1, 2, 6

    work page 2023

  8. [8]

    Causal visual-semantic correlation for zero- shot learning

    Shuhuang Chen, Dingjie Fu, Shiming Chen, Wenjin Hou, Xinge You, et al. Causal visual-semantic correlation for zero- shot learning. In ACM Multimedia 2024, 2024. 2

    work page 2024

Show all 62 references
  1. [9]

    Progressive semantic-guided vision transformer for zero-shot learning

    Shiming Chen, Wenjin Hou, Salman Khan, and Fahad Shah- baz Khan. Progressive semantic-guided vision transformer for zero-shot learning. InProceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 23964–23974, 2024

    2024

  2. [10]

    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 intelli- gence, pages 405–413, 2023. 2, 6

    2023

  3. [11]

    Latentpaint: Image inpainting in latent space with diffusion models

    Ciprian Corneanu, Raghudeep Gadde, and Aleix M Mar- tinez. Latentpaint: Image inpainting in latent space with diffusion models. In Proceedings of the IEEE/CVF Win- ter Conference on Applications of Computer Vision , pages 4334–4343, 2024. 2

    2024

  4. [12]

    An image is worth 16x16 words: Transformers for image recognition at scale

    Alexey Dosovitskiy. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929, 2020. 4

    2010 arXiv

  5. [13]

    Dream the impossible: Outlier imagination with diffusion models

    Xuefeng Du, Yiyou Sun, Jerry Zhu, and Yixuan Li. Dream the impossible: Outlier imagination with diffusion models. Advances in Neural Information Processing Systems , 36,

  6. [14]

    An image is worth one word: Personalizing text-to- image generation using textual inversion

    Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patash- nik, Amit H Bermano, Gal Chechik, and Daniel Cohen- Or. An image is worth one word: Personalizing text-to- image generation using textual inversion. arXiv preprint arXiv:2208.01618, 2022. 3

    2022 arXiv

  7. [15]

    Generative adversarial nets

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial nets. Advances in neural information processing systems, 27, 2014. 1

    2014

  8. [16]

    Is synthetic data from generative models ready for image recognition? arXiv preprint arXiv:2210.07574, 2022

    Ruifei He, Shuyang Sun, Xin Yu, Chuhui Xue, Wenqing Zhang, Philip Torr, Song Bai, and Xiaojuan Qi. Is synthetic data from generative models ready for image recognition? arXiv preprint arXiv:2210.07574, 2022. 2

    2022 arXiv

  9. [17]

    Gans trained by a two time-scale update rule converge to a local nash equilib- rium

    Martin Heusel, Hubert Ramsauer, Thomas Unterthiner, Bernhard Nessler, and Sepp Hochreiter. Gans trained by a two time-scale update rule converge to a local nash equilib- rium. Advances in neural information processing systems , 30, 2017. 7

    2017

  10. [18]

    Classifier-free diffusion guidance

    Jonathan Ho and Tim Salimans. Classifier-free diffusion guidance. arXiv preprint arXiv:2207.12598, 2022. 4

    2022 arXiv

  11. [19]

    Denoising dif- fusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising dif- fusion probabilistic models. Advances in neural information processing systems, 33:6840–6851, 2020. 4

    2020

  12. [20]

    Visual-augmented dynamic seman- tic 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 seman- tic prototype for generative zero-shot learning. In Proceed- ings of the IEEE/CVF Conference on Computer Vision and Patt...

    2024

  13. [21]

    Fine-grained generalized zero-shot learning via dense attribute-based attention

    Dat Huynh and Ehsan Elhamifar. Fine-grained generalized zero-shot learning via dense attribute-based attention. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 4483–4493, 2020. 2, 6

    2020

  14. [22]

    Transferable contrastive network for generalized zero- shot learning

    Huajie Jiang, Ruiping Wang, Shiguang Shan, and Xilin Chen. Transferable contrastive network for generalized zero- shot learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 9765–9774, 2019. 6

    2019

  15. [23]

    Kingma and M

    Diederik P. Kingma and M. Welling. Auto-encoding varia- tional bayes. In ICLR, 2014. 1

    2014

  16. [24]

    Learning to detect unseen object classes by between- class attribute transfer

    Christoph H Lampert, Hannes Nickisch, and Stefan Harmel- ing. Learning to detect unseen object classes by between- class attribute transfer. In 2009 IEEE conference on com- puter vision and pattern recognition, pages 951–958. IEEE,

    2009

  17. [25]

    Your diffusion model is secretly 9 a zero-shot classifier

    Alexander C Li, Mihir Prabhudesai, Shivam Duggal, Ellis Brown, and Deepak Pathak. Your diffusion model is secretly 9 a zero-shot classifier. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 2206–2217,

  18. [26]

    Towards understanding cross and self-attention in stable diffusion for text-guided image editing

    Bingyan Liu, Chengyu Wang, Tingfeng Cao, Kui Jia, and Jun Huang. Towards understanding cross and self-attention in stable diffusion for text-guided image editing. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 7817–7826, 2024. 2

    2024

  19. [27]

    Structure matters: Tackling the semantic discrepancy in diffusion models for image inpainting

    Haipeng Liu, Yang Wang, Biao Qian, Meng Wang, and Yong Rui. Structure matters: Tackling the semantic discrepancy in diffusion models for image inpainting. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 8038–8047, 2024. 2

    2024

  20. [28]

    Progressive semantic-visual mutual adap- tion for generalized zero-shot learning

    Man Liu, Feng Li, Chunjie Zhang, Yunchao Wei, Huihui Bai, and Yao Zhao. Progressive semantic-visual mutual adap- tion for generalized zero-shot learning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 15337–15346, 2023. 2

    2023

  21. [29]

    Distributed representations of words and phrases and their compositionality

    Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. Distributed representations of words and phrases and their compositionality. Advances in neural in- formation processing systems, 26, 2013. 2

    2013

  22. [30]

    A generative model for zero shot learn- ing using conditional variational autoencoders

    Ashish Mishra, Shiva Krishna Reddy, Anurag Mittal, and Hema A Murthy. A generative model for zero shot learn- ing using conditional variational autoencoders. In Proceed- ings of the IEEE conference on computer vision and pattern recognition workshops, pages 2188–2196, 2018. 1

    2018

  23. [31]

    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,

  24. [32]

    I2mvformer: Large language model generated multi-view document supervi- sion 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 supervi- sion for zero-shot image classification. In Proceedings of the I...

    2023

  25. [33]

    I2dformer+: Learning image to doc- ument summary attention for zero-shot image classification

    Muhammad Ferjad Naeem, Yongqin Xian, Luc Van Gool, and Federico Tombari. I2dformer+: Learning image to doc- ument summary attention for zero-shot image classification. International Journal of Computer Vision, pages 1–17, 2024. 2, 5, 6

    2024

  26. [34]

    Latent embedding feedback and discriminative features for zero-shot classification

    Sanath Narayan, Akshita Gupta, Fahad Shahbaz Khan, Cees GM Snoek, and Ling Shao. Latent embedding feedback and discriminative features for zero-shot classification. In Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XXII ...

    2020

  27. [35]

    Automated flower classification over a large number of classes

    Maria-Elena Nilsback and Andrew Zisserman. Automated flower classification over a large number of classes. In 2008 Sixth Indian conference on computer vision, graphics & im- age processing, pages 722–729. IEEE, 2008. 2, 5

    2008

  28. [36]

    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. 2, 5

    2012

  29. [37]

    Glove: Global vectors for word representation

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

    2014

  30. [38]

    Learning transferable visual models from natural language supervi- sion

    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 supervi- sion. In International conference on machine learning, ...

    2021

  31. [39]

    Hierarchical text-conditional image gener- ation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image gener- ation with clip latents. arXiv preprint arXiv:2204.06125, 1 (2):3, 2022. 2, 3

    2022 arXiv

  32. [40]

    Learning deep representations of fine-grained visual descrip- tions

    Scott Reed, Zeynep Akata, Honglak Lee, and Bernt Schiele. Learning deep representations of fine-grained visual descrip- tions. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 49–58, 2016. 1

    2016

  33. [41]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 1, 2, 3, 4, 5

    2022

  34. [42]

    Photorealistic text-to-image diffusion models with deep language understanding

    Chitwan Saharia, William Chan, Saurabh Saxena, Lala Li, Jay Whang, Emily L Denton, Kamyar Ghasemipour, Raphael Gontijo Lopes, Burcu Karagol Ayan, Tim Salimans, et al. Photorealistic text-to-image diffusion models with deep language understanding. Advances in neural information...

    2022

  35. [43]

    Term-weighting ap- proaches in automatic text retrieval

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

    1988

  36. [44]

    Generalized zero-and few-shot learning via aligned variational autoencoders

    Edgar Schonfeld, Sayna Ebrahimi, Samarth Sinha, Trevor Darrell, and Zeynep Akata. Generalized zero-and few-shot learning via aligned variational autoencoders. In Proceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8247–8255, 2019. 1, 6

    2019

  37. [45]

    Discriminative class tokens for text-to-image diffusion models

    Idan Schwartz, V ´esteinn Snæbjarnarson, Hila Chefer, Serge Belongie, Lior Wolf, and Sagie Benaim. Discriminative class tokens for text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 22725–22735, 2023. 3, 4

    2023

  38. [46]

    Zero-shot learning through cross-modal transfer

    Richard Socher, Milind Ganjoo, Christopher D Manning, and Andrew Ng. Zero-shot learning through cross-modal transfer. Advances in neural information processing systems, 26, 2013. 1, 2

    2013

  39. [47]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In International confer- ence on machine learning, pages 2256–2265. PMLR, 2015. 4

    2015

  40. [48]

    Mpnet: Masked and permuted pre-training for language understanding

    Kaitao Song, Xu Tan, Tao Qin, Jianfeng Lu, and Tie-Yan Liu. Mpnet: Masked and permuted pre-training for language understanding. Advances in neural information processing systems, 33:16857–16867, 2020. 5, 6

    2020

  41. [49]

    Learning vision from mod- 10 els rivals learning vision from data

    Yonglong Tian, Lijie Fan, Kaifeng Chen, Dina Katabi, Dilip Krishnan, and Phillip Isola. Learning vision from mod- 10 els rivals learning vision from data. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15887–15898, 2024. 3

    2024

  42. [50]

    Stablerep: Synthetic images from text-to- image models make strong visual representation learners

    Yonglong Tian, Lijie Fan, Phillip Isola, Huiwen Chang, and Dilip Krishnan. Stablerep: Synthetic images from text-to- image models make strong visual representation learners. Advances in Neural Information Processing Systems , 36,

  43. [51]

    Improving zero-shot generalization for clip with synthesized prompts

    Zhengbo Wang, Jian Liang, Ran He, Nan Xu, Zilei Wang, and Tieniu Tan. Improving zero-shot generalization for clip with synthesized prompts. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3032– 3042, 2023. 6

    2023

  44. [52]

    Caltech-ucsd birds 200

    Peter Welinder, Steve Branson, Takeshi Mita, Catherine Wah, Florian Schroff, Serge Belongie, and Pietro Perona. Caltech-ucsd birds 200. 2010. 2, 5

    2010

  45. [53]

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

    Yongqin Xian, Christoph H Lampert, Bernt Schiele, and Zeynep Akata. Zero-shot learning—a comprehensive eval- uation of the good, the bad and the ugly. IEEE transactions on pattern analysis and machine intelligence , 41(9):2251– 2265, 2018. 2, 5

    2018

  46. [54]

    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. 1, 6

    2018

  47. [55]

    f-vaegan-d2: A feature generating framework for any-shot learning

    Yongqin Xian, Saurabh Sharma, Bernt Schiele, and Zeynep Akata. f-vaegan-d2: A feature generating framework for any-shot learning. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 10275–10284, 2019. 1, 2, 6

    2019

  48. [56]

    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. Advances in Neural Information Processing Sys- tems, 33:21969–21980, 2020. 6

    2020

  49. [57]

    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. Advances in Neural Information Processing Sys- tems, 33:21969–21980, 2020. 2

    2020

  50. [58]

    Vgse: Visually-grounded semantic em- beddings for zero-shot learning

    Wenjia Xu, Yongqin Xian, Jiuniu Wang, Bernt Schiele, and Zeynep Akata. Vgse: Visually-grounded semantic em- beddings for zero-shot learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 9316–9325, 2022. 2, 5, 6

    2022

  51. [59]

    Progressive ensemble networks for zero-shot recognition

    Meng Ye and Yuhong Guo. Progressive ensemble networks for zero-shot recognition. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition , pages 11728–11736, 2019. 6

    2019

  52. [60]

    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. 2

    2024 arXiv

  53. [61]

    Learning to prompt for vision-language models

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models. In- ternational Journal of Computer Vision, 130(9):2337–2348,

  54. [2022]

    A photo of S∗ [name]

    6 11 Discriminative Image Generation with Diffusion Models for Zero-Shot Learning Supplementary Material Organization of the appendix: • Appendix A: More Implementation Details. • Appendix B: More Visualizations. A. More Implementation Details Category Discrimination Model Tra...

Pith tools

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