Generalised Zero-Shot Learning with Domain Classification in a Joint Semantic and Visual Space

Ben Harwood; Gustavo Carneiro; Michele Sasdelli; Rafael Felix

REVIEW 4 major objections 4 minor 27 references

Generalised Zero-Shot Learning with Domain Classification in a Joint Semantic and Visual Space

T0 review · 4 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A domain classifier trained in a joint visual-semantic latent space reduces seen-class bias and raises unseen accuracy in generalised zero-shot learning, achieving state-of-the-art results on CUB, SUN, AWA1, and AWA2.

desk verdict Modest incremental GZSL contribution — a domain gate inside a joint latent space — honest about its small gains, but the training/test modality mismatch for the gate deserves scrutiny. read the letter →

arxiv 1908.04930 v1 pith:PKN4LHIG submitted 2019-08-14 cs.CV

Rafael Felix , Ben Harwood , Michele Sasdelli , Gustavo Carneiro This is my paper · ORCID

classification cs.CV

keywords generalisedzero-shotlearningdomainclassificationjointlatentspacevisual-semanticembeddingseen-classbiasharmonicmeanCADA-VAEcycle-WGAN

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

This paper claims that the persistent bias of generalised zero-shot learning (GZSL) models toward seen classes can be reduced by explicitly classifying each test sample as coming from the seen or unseen domain, and combining that domain score with a class score inside a jointly learned visual-semantic latent space. The authors argue that existing GZSL methods either learn a mapping between visual and semantic spaces without checking the domain, or train separate expert classifiers, and that a single latent space shared by both modalities is a better substrate for domain classification. Across the CUB, SUN, AWA1, and AWA2 benchmarks, the proposed domain-classifier extension raises unseen-class accuracy over the CADA-VAE and cycle-WGAN baselines, for example on CUB from 48.4% to 52.9% and on SUN from 45.1% to 50.7%, with small harmonic-mean gains. If correct, the method offers a simple, model-agnostic plug-in that reduces seen-class bias and improves the practical viability of zero-shot recognition.

What carries the argument

The load-bearing object is the jointly learned latent space produced by a generative alignment model, either CADA-VAE or cycle-WGAN, in which visual features and semantic class vectors are projected into one shared embedding. On top of this latent space the paper trains a binary domain classifier that distinguishes seen from unseen samples, using semantic projections of unseen classes as the only unseen-domain training examples. The final classifier combines the class discriminator and the domain classifier as a product, so a class contributes only if its domain matches the estimated domain of the test sample. The domain classifier carries the anti-bias argument: it turns the seen/unseen distinction from an implicit property of the embedding into an explicit, learned decision.

What would settle it

Take a trained model and feed the domain classifier visual features of unseen-class test images, comparing its domain decisions against the same classifier's decisions on the semantic projections of those classes; a large drop in agreement would show the alignment assumption fails. More directly, if a held-out set of unseen-class visual features were available during evaluation, the domain classifier's own accuracy on that set would reveal whether the semantic-only training is sufficient.

Watch

Extended reading notes

Core claim

The central discovery is that a domain classifier trained on the latent space of a generative GZSL model can serve as a gate that decides whether a sample should be treated as seen or unseen, and that this gate improves classification. The proposed classifier is trained on latent projections: seen-domain samples come from both visual and semantic projections of seen classes, while unseen-domain samples come only from semantic projections of unseen classes. At test time the final score for a class y is the class-discriminator probability p(y|zx) multiplied by the domain probability f(v|zx,y), which is nonzero only when y and the estimated domain v agree. This formulation reduces the model's natural tendency to assign unseen samples to seen classes. The paper reports state-of-the-art harmonic-mean results for CUB, AWA1 and AWA2, state-of-the-art unseen accuracy for CUB, SUN, AWA1 and AWA2, and improved area under the seen/unseen accuracy curve compared with the underlying CADA-VAE and cycle-WGAN baselines.

Load-bearing premise

The domain classifier is trained on semantic projections of unseen classes but must classify visual projections of unseen images at test time, so the whole method depends on the joint latent space aligning the two modalities tightly enough that semantic projections stand in for visual ones.

Editorial extensions

If this is right

Unseen-class accuracy on CUB rises from 48.4% with the CADA-VAE baseline to 52.9% with the domain classifier, and on SUN from 45.1% to 50.7%.
The domain classifier can be grafted onto existing latent-space GZSL models without retraining their generative components, since it is trained on latent projections after the space is learned.
Because the domain gate is explicit, practitioners can calibrate the seen/unseen trade-off on a validation set, which is what the paper does before reporting AUSUC.
The approach reaches state-of-the-art harmonic mean on CUB, AWA1 and AWA2, and state-of-the-art unseen accuracy on all four benchmarks, indicating that domain gating and class discrimination are complementary rather than redundant.

Reading between the lines

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

The same domain-gating idea should transfer to latent-space models beyond CADA-VAE and cycle-WGAN, such as conditional VAE/GAN hybrids, provided their latent spaces align visual and semantic modalities; the paper only demonstrates two base models.
If visual and semantic projections of unseen classes drift apart, the domain classifier will be miscalibrated at test time; a simple diagnostic would be measuring the domain classifier's own accuracy on held-out unseen visual features, which the current evaluation only measures indirectly through final accuracy.
The paper's framing suggests that explicit domain estimation could also apply to few-shot learning, where the same seen/unseen asymmetry appears but with a handful of visual examples per novel class.
The calibration step is essential to the gains; without it the raw domain probabilities are poorly calibrated, so in deployment the validation set must resemble the test distribution.

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.

Desk Editor's Note

read the letter

Quick read of arXiv:1908.04930. The punchline: this is a small incremental contribution, but it is honest and mostly sound. The actual novelty is narrow — a binary seen/unseen domain classifier, trained inside the joint visual-semantic latent space of CADA-VAE or cycle-WGAN, whose output multiplies the class posterior. DAZSL already did domain gating; the authors know and cite it. What is new is applying the gate inside a jointly learned latent space rather than on top of separately trained classifiers. The probability decomposition in Eqs. (6)–(8) is clean, the ablation via AUSUC is appropriate, and the discussion openly admits that H-mean gains are 0.1–0.4 points. Credit where due: the paper does not oversell the per-dataset table, even if the abstract does.

The soft spots are real but not fatal. The biggest is the one the stress-test note flags: the domain classifier is trained for the unseen domain on semantic projections but evaluated on visual projections of unseen test images. CADA-VAE's alignment loss supervises the visual encoder only with seen-class images; nothing in the objective guarantees that unseen-class visual projections coincide with unseen-class semantic projections. Validation calibration is done on seen-class visual samples, so it cannot correct for this mismatch. Given that the reported gains are small, the entire effect could ride on miscalibration. A serious revision should report unseen-domain classification accuracy, visualize the latent projections, or otherwise demonstrate the alignment holds for unseen classes. Second, there is no public code, no seeds, no error bars; 0.1–0.4 H-mean differences could easily be noise. Third, tiny internal inconsistency: Sec. 5 text says CUB AUSUC for cycle-WGAN+DC improved to 0.4268, Table 3 says 0.4262. Not a big deal, but it suggests the numbers were not double-checked. The abstract's 'state-of-the-art' phrasing is also broader than Table 2 supports — on SUN, DAZSL has a higher H-mean. The claim is true if read narrowly, but the abstract does not say 'narrowly.' The reader report's comment about Eq. (8) not specifying whether the domain classifier takes y as input is fair: the conditioning says it should, but the implementation description never mentions class labels. Minor, but worth fixing.

Bottom line: the work does not change the GZSL landscape, but it is a clean plug-in that deserves a serious referee rather than a desk reject. A reviewer should ask for code, seeds, and an analysis of the unseen-domain behavior of the domain classifier. I would maybe bring it to reading group, and would cite it only if I was specifically working on domain gating in latent-space GZSL.

Referee Report

4 major / 4 minor

Summary. The paper proposes a generalized zero-shot learning (GZSL) method that augments latent-space data-augmentation models (CADA-VAE and cycle-WGAN) with a domain classifier that predicts whether a test sample belongs to the seen or unseen domain. The classification rule in Eqs. (6)-(8) multiplies a class posterior p(y|z_x) by a domain probability f(v|z_x,y), thereby down-weighting classes from the domain judged less likely. The authors evaluate on CUB, SUN, AWA1, and AWA2, reporting improved unseen-class accuracy and small H-mean gains over the base models, and they claim state-of-the-art results on several benchmarks. They also report AUSUC ablations.

Significance. If the proposal is sound, it offers a simple, model-agnostic way to mitigate the seen-class bias in latent-space GZSL models, and it is evaluated on established public benchmarks against published baselines. The paper's strengths include the use of the standardized Xian et al. protocol, comparison with a broad set of prior methods, and an ablation study. However, the reported improvements are very small (H-mean gains of 0.1-0.4 percentage points; AUSUC gains below 0.005), and the central mechanism is vulnerable to a training/test modality mismatch. The significance of the work therefore depends on whether the domain classifier remains calibrated for visual projections of unseen classes, which the paper does not establish.

major comments (4)

[Sec. 3.3, Eq. (7)-(8)] The definition of f(v|z,y) is underspecified. The domain classifier p(v|z,y) is written as a function of the candidate class y, but the paper never states how y is encoded or used by the classifier, nor whether the classifier actually receives y as an input. The training description in Sec. 3.3 only says that samples are drawn from the latent space for the seen and unseen domains; it does not describe the label conditioning. Because Eq. (7) is the final classification rule, this omission makes the method non-reproducible as written and should be clarified with the exact input representation and architecture.
[Sec. 3.3 and Sec. 4.3] The domain classifier is trained for the unseen domain using semantic projections in the latent space, but at test time it is applied to visual projections Encoder_x(x) of unseen images. The visual encoder is trained only on seen-class images, and the alignment losses in Eqs. (4)-(5) do not by themselves guarantee that semantic projections of unseen classes and visual projections of unseen images occupy the same region of the joint latent space. Calibration on the validation set, which contains only seen-class visual samples, cannot detect miscalibration on unseen visual projections. The paper should provide direct evidence of cross-modal alignment for unseen classes, for example by measuring distances between semantic and visual projections of unseen classes in the latent space, or by evaluating the domain classifier's accuracy on held-out unseen visual samples if any are available.
[Table 2 and Section 6] The conclusion states that the proposed approach achieves state-of-the-art unseen accuracy for AWA1, but Table 2 shows f-CLSWGAN with Y_U = 57.9% versus CADA-VAE + DC with Y_U = 57.5% on AWA1. This contradicts the stated claim. The authors should correct the claim or explain the basis for it (e.g., if they are comparing only to the reproduced CADA-VAE baseline).
[Tables 2 and 3] The reported improvements over the CADA-VAE baseline are very small: H-mean gains are 0.1-0.4 percentage points and AUSUC differences are at most 0.005. No error bars, standard deviations, or significance tests are reported, and the paper does not state how many random seeds were used. Given the magnitude of the differences, the central empirical claim that the domain classifier helps is not statistically supported. The authors should report multiple runs with variance and, if possible, a paired significance test.

minor comments (4)

[Eq. (4)] The notation L is used in the cross-reconstruction term without being defined; the number of modalities and the summation indices should be stated explicitly.
[Eq. (5)] The WGAN loss term L_WGAN is referenced but not defined; a brief definition or citation to the exact formulation would improve clarity.
[Sec. 5] The text describes H-mean improvements as "0.2%, 0.1%, 0.1%, and 0.4%" but appears to mean percentage points; the wording is ambiguous and should be clarified.
[Sec. 1] The abstract and introduction report "improvements of 4.5%, 5.6%, 2.5%, 1.5%" in unseen accuracy; these should be labeled as absolute percentage points, not relative improvements, to avoid misinterpretation.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the domain classifier is trained on latent-space samples and evaluated on held-out benchmarks, and the only self-citation is a published baseline rather than a load-bearing premise.

full rationale

The paper's derivation chain is a probabilistic decomposition, not a self-referential fit. Equation (6)-(8) define the GZSL classifier as f(y|x) = sum_v p(y|z_x) f(v|z_x,y), with f(v|z_x,y) equal to the learned domain probability p(v|z_x,y) when the domain matches and zero otherwise. This is a construction based on learned components, and the final claim is measured on held-out test images from external benchmarks (CUB, SUN, AWA1, AWA2) against published baselines. The domain classifier is trained using latent-space samples: seen samples from visual and semantic projections, and unseen samples from semantic projections only (Sec. 3.3). The fact that at test time it receives visual projections of unseen images is a train/test modality-alignment assumption and a potential calibration risk, but it is not a circular reduction: the unseen accuracy and H-mean are not defined in terms of the domain classifier's training outputs, and no parameter is fitted to the reported metrics. The only self-citation is the cycle-WGAN baseline [3], a published ECCV 2018 method by the same first author; it is used as a comparison backbone with reproduced results, not as the justification for the claimed improvement. The calibration of the domain classifier on the validation set is a standard post-hoc step and does not redefine the evaluation metric. Therefore, no load-bearing step in the paper reduces by construction to its own inputs.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central model is a supervised combination of two learned networks. The main premise that is not paid for by the paper is that semantic projections of unseen classes, used to train the domain classifier, are a faithful proxy for visual projections of unseen images at test time. The paper also trusts the standard GZSL benchmark protocol from Xian et al. [1] and fixed feature extractors.

free parameters (3)

Domain classifier network weights = not reported
The binary domain classifier is trained on the latent space; its weights are learned from seen visual and semantic projections and unseen semantic projections. The paper does not report their values or architecture details.
Domain classifier calibration parameters = not reported
Section 4.3 states the domain classifier output 'tends not to be well calibrated' and is calibrated using the validation set, introducing fitted offset or scale parameters that are not reported.
Latent space hyperparameters (gamma, delta, hidden sizes) = from [4] and [3] (gamma 0.044, delta 0.0026 scheduling)
The latent-space learning uses hyperparameters copied from CADA-VAE and cycle-WGAN; they are chosen by the original authors and not re-tuned in this paper, but they affect the results.

assumptions (4)

domain assumption Semantic projections of unseen classes are a valid proxy for visual projections of unseen classes in the joint latent space.
The domain classifier is trained with unseen-domain samples taken from semantic projections only (Sec. 3.3), while at test time it must classify visual projections of unseen images.
ad hoc to paper The encoder posterior p(z|x) is a delta function at zx = Encoderx(x).
Eq. (7) drops the integration over z and assumes certainty in the encoding, ignoring encoder variance.
domain assumption Evaluation protocol and feature extractors from [1] are correct and comparable across methods.
The paper adopts Xian et al.'s splits, ResNet-101 features, and semantic features; all comparisons inherit these choices.
domain assumption The base latent-space models (CADA-VAE, cycle-WGAN) are faithfully reproduced.
The paper reproduces baseline numbers marked with an asterisk, but no code is released to verify the reproduction.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Generalised Zero-Shot Learning with Domain Classification in a Joint Semantic and Visual Space." pith.science (2026). https://pith.science/paper/PKN4LHIG

@misc{pith2026190804930,
  author       = {Pith},
  title        = {Pith review of: Generalised Zero-Shot Learning with Domain Classification in a Joint Semantic and Visual Space},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PKN4LHIG}},
  note         = {Machine review of arXiv:1908.04930}
}

read the original abstract

Generalised zero-shot learning (GZSL) is a classification problem where the learning stage relies on a set of seen visual classes and the inference stage aims to identify both the seen visual classes and a new set of unseen visual classes. Critically, both the learning and inference stages can leverage a semantic representation that is available for the seen and unseen classes. Most state-of-the-art GZSL approaches rely on a mapping between latent visual and semantic spaces without considering if a particular sample belongs to the set of seen or unseen classes. In this paper, we propose a novel GZSL method that learns a joint latent representation that combines both visual and semantic information. This mitigates the need for learning a mapping between the two spaces. Our method also introduces a domain classification that estimates whether a sample belongs to a seen or an unseen class. Our classifier then combines a class discriminator with this domain classifier with the goal of reducing the natural bias that GZSL approaches have toward the seen classes. Experiments show that our method achieves state-of-the-art results in terms of harmonic mean, the area under the seen and unseen curve and unseen classification accuracy on public GZSL benchmark data sets. Our code will be available upon acceptance of this paper.

Figures

Figures reproduced from arXiv: 1908.04930 by the authors.

**Figure 3.** Depiction of the cycle-WGAN method [3]. This method encodes the semantic space into a latent visual space. The decoder produces semantic vectors that are used to regularise the learning process. where we assume from (6) that p(z | x) is a delta function at zx = Encoderx(x). The term p(y | zx) in (7) is represented by a simple deep learning classifier with softmax activation. We define the function f(.) in (7) by f(v… view at source ↗

**Figure 4.** Example of two classes that are visually [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 23 canonical work pages

[1]

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

Y . Xian, C. H. Lampert, B. Schiele, and Z. Akata, “Zero-shot learning - A comprehensive evaluation of the good, the bad and the ugly,” CoRR, vol. abs/1707.00600, 2017

arXiv 2017
[2]

Devise: A deep visual- semantic embedding model,

A. Frome, G. S. Corrado, J. Shlens, S. Bengio, J. Dean, T. Mikolov et al. , “Devise: A deep visual- semantic embedding model,” in Advances in neural information processing systems, 2013, pp. 2121–2129

work page 2013
[3]

Multi-modal cycle-consistent generalized zero-shot learning,

R. Felix, B. V . Kumar, I. Reid, and G. Carneiro, “Multi-modal cycle-consistent generalized zero-shot learning,” in European Conference on Computer Vi- sion. Springer, 2018, pp. 21–37

work page 2018
[4]

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

E. Schonfeld, S. Ebrahimi, S. Sinha, T. Darrell, and Z. Akata, “Generalized zero-and few-shot learning via aligned variational autoencoders,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 8247–8255

work page 2019
[5]

Adaptive conﬁdence smoothing for generalized zero-shot learning,

Y . Atzmon and G. Chechik, “Adaptive conﬁdence smoothing for generalized zero-shot learning,” in Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 11 671–11 680

work page 2019
[6]

Zero-shot learning through cross-modal transfer,

R. Socher, M. Ganjoo, C. D. Manning, and A. Ng, “Zero-shot learning through cross-modal transfer,” in Advances in Neural Information Processing Systems , 2013, pp. 935–943

work page 2013
[7]

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

C. H. Lampert, H. Nickisch, and S. Harmeling, “Learning to detect unseen object classes by between- class attribute transfer,” in Computer Vision and Pat- tern Recognition, 2009. CVPR 2009. IEEE Conference on, June 2009, pp. 951–958

work page 2009
[8]

Caltech-ucsd birds 200,

P. Welinder, S. Branson, T. Mita, C. Wah, F. Schroff, S. Belongie, and P. Perona, “Caltech-ucsd birds 200,” 2010

2010

Show all 27 references

[9]

An empirical study and analysis of generalized zero- shot learning for object recognition in the wild,

W.-L. Chao, S. Changpinyo, B. Gong, and F. Sha, “An empirical study and analysis of generalized zero- shot learning for object recognition in the wild,” in European Conference on Computer Vision. Springer, 2016, pp. 52–68

2016
[10]

Attribute-based classiﬁcation for zero-shot visual ob- ject categorization,

C. H. Lampert, H. Nickisch, and S. Harmeling, “Attribute-based classiﬁcation for zero-shot visual ob- ject categorization,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 36, no. 3, pp. 453–465, 2014

2014
[11]

Feature generating networks for zero-shot learning,

Y . Xian, T. Lorenz, B. Schiele, and Z. Akata, “Feature generating networks for zero-shot learning,” arXiv, 2017

2017
[12]

Synthesized classiﬁers for zero-shot learning,

S. Changpinyo, W.-L. Chao, B. Gong, and F. Sha, “Synthesized classiﬁers for zero-shot learning,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 5327–5336

2016
[13]

Gener- alized zero-shot learning via synthesized examples,

V . K. Verma, G. Arora, A. Mishra, and P. Rai, “Gener- alized zero-shot learning via synthesized examples,” in The IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018

2018
[14]

Deep residual learning for image recognition,

K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778

2016
[15]

Semantic autoen- coder for zero-shot learning,

T. X. Elyor Kodirov and S. Gong, “Semantic autoen- coder for zero-shot learning,” IEEE CVPR 2017 , July 2017

2017
[16]

A generative model for zero shot learning using conditional variational autoencoders,

A. Mishra, S. Krishna Reddy, A. Mittal, and H. A. Murthy, “A generative model for zero shot learning using conditional variational autoencoders,” in Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2018, pp. 2188– 2196

2018
[17]

A generative adversarial approach for zero- shot learning from noisy texts,

Y . Zhu, M. Elhoseiny, B. Liu, X. Peng, and A. El- gammal, “A generative adversarial approach for zero- shot learning from noisy texts,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 1004–1013

2018
[18]

Sun database: Large-scale scene recognition from abbey to zoo,

J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Tor- ralba, “Sun database: Large-scale scene recognition from abbey to zoo,” in Computer vision and pattern recognition (CVPR), 2010 IEEE conference on. IEEE, 2010, pp. 3485–3492

2010
[19]

Learning deep representations of ﬁne-grained visual descrip- tions,

S. Reed, Z. Akata, H. Lee, and B. Schiele, “Learning deep representations of ﬁne-grained visual descrip- tions,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2016, pp. 49–58

2016
[20]

Adam: A method for stochastic optimization,

D. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv, 2014

2014
[21]

Rectiﬁer nonlinearities improve neural network acoustic mod- els,

A. L. Maas, A. Y . Hannun, and A. Y . Ng, “Rectiﬁer nonlinearities improve neural network acoustic mod- els,” in Proc. icml, vol. 30, no. 1, 2013, p. 3

2013
[22]

Rectiﬁed linear units improve restricted boltzmann machines,

V . Nair and G. E. Hinton, “Rectiﬁed linear units improve restricted boltzmann machines,” in Proceed- ings of the 27th international conference on machine learning (ICML-10), 2010, pp. 807–814

2010
[23]

Concrete dropout,

Y . Gal, J. Hron, and A. Kendall, “Concrete dropout,” in Advances in Neural Information Processing Systems , 2017, pp. 3581–3590

2017
[24]

Evaluation of output embeddings for ﬁne-grained image classiﬁcation,

Z. Akata, S. Reed, D. Walter, H. Lee, and B. Schiele, “Evaluation of output embeddings for ﬁne-grained image classiﬁcation,” in Proceedings of the IEEE Con- ference on Computer Vision and Pattern Recognition , 2015, pp. 2927–2936

2015
[25]

Label-embedding for image classiﬁcation,

Z. Akata, F. Perronnin, Z. Harchaoui, and C. Schmid, “Label-embedding for image classiﬁcation,” IEEE transactions on pattern analysis and machine intel- ligence, vol. 38, no. 7, pp. 1425–1438, 2016

2016
[26]

Latent embeddings for zero-shot classiﬁcation,

Y . Xian, Z. Akata, G. Sharma, Q. Nguyen, M. Hein, and B. Schiele, “Latent embeddings for zero-shot classiﬁcation,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , 2016, pp. 69–77

2016
[27]

An embarrassingly simple approach to zero-shot learning,

B. Romera-Paredes and P. Torr, “An embarrassingly simple approach to zero-shot learning,” in Interna- tional Conference on Machine Learning , 2015, pp. 2152–2161

2015

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

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

Y . Xian, C. H. Lampert, B. Schiele, and Z. Akata, “Zero-shot learning - A comprehensive evaluation of the good, the bad and the ugly,” CoRR, vol. abs/1707.00600, 2017

arXiv 2017

[2] [2]

Devise: A deep visual- semantic embedding model,

A. Frome, G. S. Corrado, J. Shlens, S. Bengio, J. Dean, T. Mikolov et al. , “Devise: A deep visual- semantic embedding model,” in Advances in neural information processing systems, 2013, pp. 2121–2129

work page 2013

[3] [3]

Multi-modal cycle-consistent generalized zero-shot learning,

R. Felix, B. V . Kumar, I. Reid, and G. Carneiro, “Multi-modal cycle-consistent generalized zero-shot learning,” in European Conference on Computer Vi- sion. Springer, 2018, pp. 21–37

work page 2018

[4] [4]

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

E. Schonfeld, S. Ebrahimi, S. Sinha, T. Darrell, and Z. Akata, “Generalized zero-and few-shot learning via aligned variational autoencoders,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 8247–8255

work page 2019

[5] [5]

Adaptive conﬁdence smoothing for generalized zero-shot learning,

Y . Atzmon and G. Chechik, “Adaptive conﬁdence smoothing for generalized zero-shot learning,” in Pro- ceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 11 671–11 680

work page 2019

[6] [6]

Zero-shot learning through cross-modal transfer,

R. Socher, M. Ganjoo, C. D. Manning, and A. Ng, “Zero-shot learning through cross-modal transfer,” in Advances in Neural Information Processing Systems , 2013, pp. 935–943

work page 2013

[7] [7]

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

C. H. Lampert, H. Nickisch, and S. Harmeling, “Learning to detect unseen object classes by between- class attribute transfer,” in Computer Vision and Pat- tern Recognition, 2009. CVPR 2009. IEEE Conference on, June 2009, pp. 951–958

work page 2009

[8] [8]

Caltech-ucsd birds 200,

P. Welinder, S. Branson, T. Mita, C. Wah, F. Schroff, S. Belongie, and P. Perona, “Caltech-ucsd birds 200,” 2010

2010