Academia.eduAcademia.edu

Outline

Title
Abstract
Figures
Introduction
Inductive and Semantic Transductive GZSL Methods
Graph-Based Methods
Meta Learning-Based Methods
Attention-Based Methods
Compositional Learning-Based Methods
Bidirectional Learning Methods
Autoencoder-Based Methods
Other Methods
Transductive GZSL Methods
Embedding-Based Methods
Generative-Based Methods
Discussions and Conclusions
References

A Review of Generalized Zero-Shot Learning Methods

Profile image of Farhad PourpanahFarhad Pourpanah

2020

Abstract

Generalized zero-shot learning (GZSL) aims to train a model for classifying data samples under the condition that some output classes are unknown during supervised learning. To address this challenging task, GZSL leverages semantic information of both seen (source) and unseen (target) classes to bridge the gap between both seen and unseen classes. Since its introduction, many GZSL models have been formulated. In this review paper, we present a comprehensive review of GZSL. Firstly, we provide an overview of GZSL including the problems and challenging issues. Then, we introduce a hierarchical categorization of the GZSL methods and discuss the representative methods of each category. In addition, we discuss several research directions for future studies.

Figures (20)
Index Terms—Generalized zero shot learning, deep learning, semantic embedding, generative adversarial networks, variational auto-encoders Fig. 1: A schematic diagram of zero-shot learning (ZSL) versus generalized zero-shot learning (GZSL). Assume that the seen class set includes Otter and Tiger, while the unseen class set contains Polar bear and Zebra. During the training stage (a), both GZSL and ZSL have access to the samples and semantic representations of the seen class set. During the test stage, (b) ZSL can only recognize samples from the unseen class set, e.g., Polar bear and Zebra, while (c) GZSL is able to recognize samples from both sets, e.g., Otter, Tiger, Polar bear and Zebra.
Index Terms—Generalized zero shot learning, deep learning, semantic embedding, generative adversarial networks, variational auto-encoders Fig. 1: A schematic diagram of zero-shot learning (ZSL) versus generalized zero-shot learning (GZSL). Assume that the seen class set includes Otter and Tiger, while the unseen class set contains Polar bear and Zebra. During the training stage (a), both GZSL and ZSL have access to the samples and semantic representations of the seen class set. During the test stage, (b) ZSL can only recognize samples from the unseen class set, e.g., Polar bear and Zebra, while (c) GZSL is able to recognize samples from both sets, e.g., Otter, Tiger, Polar bear and Zebra.
Fig. 2: A schematic view of the main difference between Transductive GZSL, Transductive semantic GZSL and In- ductive GZSL.
Fig. 2: A schematic view of the main difference between Transductive GZSL, Transductive semantic GZSL and In- ductive GZSL.
2.3. Embedding Spaces
2.3. Embedding Spaces
Fig. 5: A schematic view of the bias concerning seen classes (source) in the semantic embedding space (adapted from [60]).
Fig. 5: A schematic view of the bias concerning seen classes (source) in the semantic embedding space (adapted from [60]).
Fig. 6: Embedding-based versus generative-based methods. The embedding based methods (a) lean an embedding space to project the visual and semantic features of seen classes into a common space. Then they measure the similarity level to predict novel classes. The generative-based methods (b) learn a generative model based on samples of seen classes conditioned on their semantic features. Then, the learned model is used to generate visual features for unseen classes using the semantic features of unseen classes and convert GZSL into a supervised learning problem.
Fig. 6: Embedding-based versus generative-based methods. The embedding based methods (a) lean an embedding space to project the visual and semantic features of seen classes into a common space. Then they measure the similarity level to predict novel classes. The generative-based methods (b) learn a generative model based on samples of seen classes conditioned on their semantic features. Then, the learned model is used to generate visual features for unseen classes using the semantic features of unseen classes and convert GZSL into a supervised learning problem.
Fig. 7: The taxonomy of GZSL models.
Fig. 7: The taxonomy of GZSL models.
Image Feature Space Shared Reconstruction Graph semantic Embedding space Fig. 8: A schematic view of SRG [52]. Firstly, the image prototype f;, and semantic prototype e, of each class are reconstructed using the cluster center and (7), respectively. Then, it learns the shared reconstruction coefficients between two spaces. The aim is to force same classes share same reconstruction coefficients.Finally, the learned SRG is used to synthesize class prototypes for unseen classes to perform prediction.
Image Feature Space Shared Reconstruction Graph semantic Embedding space Fig. 8: A schematic view of SRG [52]. Firstly, the image prototype f;, and semantic prototype e, of each class are reconstructed using the cluster center and (7), respectively. Then, it learns the shared reconstruction coefficients between two spaces. The aim is to force same classes share same reconstruction coefficients.Finally, the learned SRG is used to synthesize class prototypes for unseen classes to perform prediction.
Fig. 9: A schematic of DAZLE [71]. After extracting image features of the R regions, the attention features of all attribute are computed using a dense attention mechanism. Then, the attention features are aligned with the attribute semant vectors, in order to compute the score of attributes in the image.
Fig. 9: A schematic of DAZLE [71]. After extracting image features of the R regions, the attention features of all attribute are computed using a dense attention mechanism. Then, the attention features are aligned with the attribute semant vectors, in order to compute the score of attributes in the image.
Fig. 10: A schematic view of JIL [114]. Firstly, it jointly projects visual and semantic spaces into a shared subspace. Then, reconstructs each space to learn more generalizable projections for GZSL.
Fig. 10: A schematic view of JIL [114]. Firstly, it jointly projects visual and semantic spaces into a shared subspace. Then, reconstructs each space to learn more generalizable projections for GZSL.
ait “iat teal Fig. 12: A schematic view of CDL [125]. After obtaining the class prototypes in both spaces, a coupled dictionary learn- ing framework is developed to align the visual-semantic structure. Finally, a domain adaptation is proposed to learn the unseen class prototypes in the visual space to assist the structure learning process.
ait “iat teal Fig. 12: A schematic view of CDL [125]. After obtaining the class prototypes in both spaces, a coupled dictionary learn- ing framework is developed to align the visual-semantic structure. Finally, a domain adaptation is proposed to learn the unseen class prototypes in the visual space to assist the structure learning process.
\BLE 1: A summary of embedding-based methods (“S”, “V” and “L” represent the semantic, visual and latent embeddins ace, respectively, and “-” indicates the item is not reported by the corresponding study). projection matrix to that space. Specifically, it fixes the labels of the seen classes and uses the attributes to map the label of the unseen classes into the label space. Liu et al. [128] used a graph attention network [130] to generate optimized attribute vector for each class. generated by data-driven clustering, across the seen and unseen class domains to tune two domains. Jin et al. [126] projected the semantic vector of each class to a high-order attribute space by applying the Gaussian random projection. The attributing label space (ALS) [127] defines a discrimina- tive label space and projects the visual features via a linear
\BLE 1: A summary of embedding-based methods (“S”, “V” and “L” represent the semantic, visual and latent embeddins ace, respectively, and “-” indicates the item is not reported by the corresponding study). projection matrix to that space. Specifically, it fixes the labels of the seen classes and uses the attributes to map the label of the unseen classes into the label space. Liu et al. [128] used a graph attention network [130] to generate optimized attribute vector for each class. generated by data-driven clustering, across the seen and unseen class domains to tune two domains. Jin et al. [126] projected the semantic vector of each class to a high-order attribute space by applying the Gaussian random projection. The attributing label space (ALS) [127] defines a discrimina- tive label space and projects the visual features via a linear
Fig. 13: A general view of the f-CLSWGAN model [138]). It learns a generative model based on the visual features of seen classes conditioned on their corresponding semantic representations. It also uses classification loss to generate more discriminative visual features.
Fig. 13: A general view of the f-CLSWGAN model [138]). It learns a generative model based on the visual features of seen classes conditioned on their corresponding semantic representations. It also uses classification loss to generate more discriminative visual features.
Fig. 14: A schematic view of CVAE-ZSL [164]. After concate- nating the visual features and the semantic representations, it is passed through dense, dropout and dense layers. Then, another dense layer is used to output yw, and >>. Next, a z is sampled from N(fi2,, Uz,) and projected to the image space for reconstruction.
Fig. 14: A schematic view of CVAE-ZSL [164]. After concate- nating the visual features and the semantic representations, it is passed through dense, dropout and dense layers. Then, another dense layer is used to output yw, and >>. Next, a z is sampled from N(fi2,, Uz,) and projected to the image space for reconstruction.
Fig. 15: A general view of Zero-VAE-GAN [166]. After concatenating the visual features and semantic representations of seen class, an encoder is used to project the real features into a latent space. Then, a generator is used to decode the latent representation into the feature space for reconstruction. In addition, the adversarial discriminator and categorizer are, respectively, used to generate more discriminative and similar visual features to the real ones. For transductive setting, Zero- VAE-GAN uses self-training strategy to tune the model using the unlabeled visual features and semantic representations of unseen classes.
Fig. 15: A general view of Zero-VAE-GAN [166]. After concatenating the visual features and semantic representations of seen class, an encoder is used to project the real features into a latent space. Then, a generator is used to decode the latent representation into the feature space for reconstruction. In addition, the adversarial discriminator and categorizer are, respectively, used to generate more discriminative and similar visual features to the real ones. For transductive setting, Zero- VAE-GAN uses self-training strategy to tune the model using the unlabeled visual features and semantic representations of unseen classes.
Fig. 16: A schematic view of the exemplar synthesis (EXEM) [66]. A kernel-based regressor ~)(.) is learned based on the visual features and semantic representations of seen classes so that the semantic representation of each seen class c can recognize its visual exemplar (i.e., center) z.. Then is used to predict visual features of the unseen classes by a nearest neighbor search.
Fig. 16: A schematic view of the exemplar synthesis (EXEM) [66]. A kernel-based regressor ~)(.) is learned based on the visual features and semantic representations of seen classes so that the semantic representation of each seen class c can recognize its visual exemplar (i.e., center) z.. Then is used to predict visual features of the unseen classes by a nearest neighbor search.
TABLE 2: A summary of generative-based methods (“-” indicates the item is not reported by the corresponding study).14
TABLE 2: A summary of generative-based methods (“-” indicates the item is not reported by the corresponding study).14
Fig. 17: A general view of OFSL [187]. After projecting the visual features of the target (seen) classes into a number of fixed points in the semantic space, a bias loss function is used to map the visual features of the target (unseen) classes into other points.
Fig. 17: A general view of OFSL [187]. After projecting the visual features of the target (seen) classes into a number of fixed points in the semantic space, a bias loss function is used to map the visual features of the target (unseen) classes into other points.
ABLE 3: A summary of transductive-based methods (“S”, “V” and “L” represent the semantic, visual and later mbedding spaces, respectively, and “-” indicates the score is not reported by the corresponding study).
ABLE 3: A summary of transductive-based methods (“S”, “V” and “L” represent the semantic, visual and later mbedding spaces, respectively, and “-” indicates the score is not reported by the corresponding study).

Related papers

Zero-Shot Learning with Missing Attributes using Semantic Correlations
Kunal Chaturvedi

2021 International Joint Conference on Neural Networks (IJCNN), 2021

Zero-shot learning (ZSL) aims to recognize instances belonging to unseen categories which are not available at training time. Previous ZSL models learn a projection function from the visual feature space to a semantic space which contains a description of the categories. The semantic attributes are often correlated with each other at the semantic space and it is not appropriate to learn them independently. Existing ZSL methods are designed to work on complete descriptions of the semantic attributes. However, because these attributes are human-designed values, they might be incomplete or contains noisy values which may affect the recognition performance of many existing ZSL models. This paper proposes a novel zero-shot learning approach (ZSL-MSA) to handle missing and noisy semantic attributes during the training process. Significantly, the proposed method learns a supplementary attribute matrix by exploiting the attribute correlation. The proposed method also learns the relevant feature coefficients in the projection matrix to identify the correlated attribute space. Th proposed method also adopts l1 regularization norm to select the relevant sparse features. A constrained optimization function is formulated and solved using the accelerated proximal gradient method. Extensive experiments on three benchmark datasets using ZSL and generalized ZSL demonstrate the effectiveness of the proposed method.

View PDFchevron_right
Exploring the Limits of Zero-Shot Learning - How Low Can You Go?
karan sharma

2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

Standard zero-shot learning (ZSL) methods use a large number of seen categories to predict very few unseen categories while maintaining unified data splits and evaluation metrics. This has enabled the research community to advance notably towards formulating a standard benchmark ZSL algorithm. However, the most substantial impact of ZSL lies in enabling the prediction of a large number of unseen categories from very few seen categories within a specific domain. This permits the collection and annotation of training data for only a few previously seen categories, thereby significantly mitigating the training data collection and annotation process. We address the difficult problem of predicting a large number of unseen object categories from very few previously seen categories and propose a framework that enables us to examine the limits of inferring several unseen object categories from very few previously seen object categories, i.e., the limits of ZSL. We examine the functional dependence of the classification accuracy of unseen object classes on the number and types of previously seen classes and determine the minimum number and types of previously seen classes required to achieve a prespecified classification accuracy for the unseen classes on three standard ZSL data sets. An experimental comparison of the proposed framework to a prominent ZSL technique on these data sets shows that the proposed framework achieves higher classification accuracy on average while providing valuable insights into the unseen class inference process.

View PDFchevron_right
Unified Generator-Classifier for Efficient Zero-Shot Learning
Ayyappa Kumar Pambala

2019

Generative models have achieved state-of-the-art performance for the zero-shot learning problem, but they require re-training the classifier every time a new object category is encountered. The traditional semantic embedding approaches, though very elegant, usually do not perform at par with their generative counterparts. In this work, we propose an unified framework termed GenClass, which integrates the generator with the classifier for efficient zero-shot learning, thus combining the representative power of the generative approaches and the elegance of the embedding approaches. End-to-end training of the unified framework not only eliminates the requirement of additional classifier for new object categories as in the generative approaches, but also facilitates the generation of more discriminative and useful features. Extensive evaluation on three standard zero-shot object classification datasets, namely AWA, CUB and SUN shows the effectiveness of the proposed approach. The approa...

View PDFchevron_right
Zero-shot Learning with Class Description Regularization
Marcus Brubaker

arXiv (Cornell University), 2021

The purpose of generative Zero-shot learning (ZSL) is to learning from seen classes, transfer the learned knowledge, and create samples of unseen classes from the description of these unseen categories. To achieve better ZSL accuracies, models need to better understand the descriptions of unseen classes. We introduce a novel form of regularization that encourages generative ZSL models to pay more attention to the description of each category. Our empirical results demonstrate improvements over the performance of multiple state-of-the-art models on the task of generalized zeroshot recognition and classification when trained on textual description-based datasets like CUB and NABirds and attribute-based datasets like AWA2, aPY and SUN. Code is available at https://github.com/shayan-kousha/DGRZSL

View PDFchevron_right
A Meta-Learning Framework for Generalized Zero-Shot Learning
Dhanajit Brahma

ArXiv, 2019

Learning to classify unseen class samples at test time is popularly referred to as zero-shot learning (ZSL). If test samples can be from training (seen) as well as unseen classes, it is a more challenging problem due to the existence of strong bias towards seen classes. This problem is generally known as \emph{generalized} zero-shot learning (GZSL). Thanks to the recent advances in generative models such as VAEs and GANs, sample synthesis based approaches have gained considerable attention for solving this problem. These approaches are able to handle the problem of class bias by synthesizing unseen class samples. However, these ZSL/GZSL models suffer due to the following key limitations: $(i)$ Their training stage learns a class-conditioned generator using only \emph{seen} class data and the training stage does not \emph{explicitly} learn to generate the unseen class samples; $(ii)$ They do not learn a generic optimal parameter which can easily generalize for both seen and unseen cl...

View PDFchevron_right
A Simple Exponential Family Framework for Zero-Shot Learning
Piyush Rai

Machine Learning and Knowledge Discovery in Databases, 2017

We present a simple generative framework for learning to predict previously unseen classes, based on estimating class-attribute-gated class-conditional distributions. We model each class-conditional distribution as an exponential family distribution and the parameters of the distribution of each seen/unseen class are defined as functions of the respective observed class attributes. These functions can be learned using only the seen class data and can be used to predict the parameters of the class-conditional distribution of each unseen class. Unlike most existing methods for zero-shot learning that represent classes as fixed embeddings in some vector space, our generative model naturally represents each class as a probability distribution. It is simple to implement and also allows leveraging additional unlabeled data from unseen classes to improve the estimates of their class-conditional distributions using transductive/semi-supervised learning. Moreover, it extends seamlessly to few-shot learning by easily updating these distributions when provided with a small number of additional labelled examples from unseen classes. Through a comprehensive set of experiments on several benchmark data sets, we demonstrate the efficacy of our framework.

View PDFchevron_right
Leveraging Seen and Unseen Semantic Relationships for Generative Zero-Shot Learning
maunil vyas

ArXiv, 2020

Zero-shot learning (ZSL) addresses the unseen class recognition problem by leveraging semantic information to transfer knowledge from seen classes to unseen classes. Generative models synthesize the unseen visual features and convert ZSL into a classical supervised learning problem. These generative models are trained using the seen classes and are expected to implicitly transfer the knowledge from seen to unseen classes. However, their performance is stymied by overfitting, which leads to substandard performance on Generalized Zero-Shot learning (GZSL). To address this concern, we propose the novel LsrGAN, a generative model that Leverages the Semantic Relationship between seen and unseen categories and explicitly performs knowledge transfer by incorporating a novel Semantic Regularized Loss (SR-Loss). The SR-loss guides the LsrGAN to generate visual features that mirror the semantic relationships between seen and unseen classes. Experiments on seven benchmark datasets, including t...

View PDFchevron_right
Zero-shot learning with semantic output codes
Dean Pomerleau, Tom Mitchell

2009

We consider the problem of zero-shot learning, where the goal is to learn a classifier f : X → Y that must predict novel values of Y that were omitted from the training set. To achieve this, we define the notion of a semantic output code classifier (SOC) which utilizes a knowledge base of semantic properties of Y to extrapolate to novel classes. We provide a formalism for this type of classifier and study its theoretical properties in a PAC framework, showing conditions under which the classifier can accurately predict novel classes. As a case study, we build a SOC classifier for a neural decoding task and show that it can often predict words that people are thinking about from functional magnetic resonance images (fMRI) of their neural activity, even without training examples for those words.

View PDFchevron_right
Learning without Seeing nor Knowing: Towards Open Zero-Shot Learning
federico marmoreo

arXiv (Cornell University), 2021

In Generalized Zero-Shot Learning (GZSL), unseen categories (for which no visual data are available at training time) can be predicted by leveraging their class embeddings (e.g., a list of attributes describing them) together with a complementary pool of seen classes (paired with both visual data and class embeddings). Despite GZSL is arguably challenging, we posit that knowing in advance the class embeddings, especially for unseen categories, is an actual limit of the applicability of GZSL towards real-world scenarios. To relax this assumption, we propose Open Zero-Shot Learning (OZSL) to extend GZSL towards the openworld settings. We formalize OZSL as the problem of recognizing seen and unseen classes (as in GZSL) while also rejecting instances from unknown categories, for which neither visual data nor class embeddings are provided. We formalize the OZSL problem introducing evaluation protocols, error metrics and benchmark datasets. We also suggest to tackle the OZSL problem by proposing the idea of performing unknown feature generation (instead of only unseen features generation as done in GZSL). We achieve this by optimizing a generative process to sample unknown class embeddings as complementary to the seen and the unseen. We intend these results to be the ground to foster future research, extending the standard closed-world zeroshot learning (GZSL) with the novel open-world counterpart (OZSL).

View PDFchevron_right
Transductive Multi-class and Multi-label Zero-shot Learning
Yanwei Fu

Cornell University - arXiv, 2015

View PDFchevron_right

References (237)

  1. J. Deng, W. Dong, R. Socher, L. Li, Kai Li, and Li Fei-Fei, "Ima- geNet: A large-scale hierarchical image database," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 248-255.
  2. S. Liu, M. Long, J. Wang, and M. I. Jordan, "Generalized zero-shot learning with deep calibration network," in Advances in Neural Information Processing Systems, 2018, pp. 2005-2015.
  3. W. Wang, V. W. Zheng, H. Yu, and C. Miao, "A survey of zero-shot learning: Settings, methods, and applications," ACM Transactions on Intelligent Systems and Technology, vol. 10, no. 2, pp. 1-37, 2019.
  4. X. Wang, Y. Zhao, and F. Pourpanah, "Recent advances in deep learning," International Journal of Machine Learning and Cybernetics, vol. 11, pp. 747-750, 2020.
  5. H. Kabir, M. Abdar, S. M. J. Jalali, A. Khosravi, A. F. Atiya, S. Na- havandi, and D. Srinivasan, "Spinalnet: Deep neural network with gradual input," arXiv preprint:2007.03347, 2020.
  6. M. Abdar, F. Pourpanah, S. Hussain, D. Rezazadegan, L. Liu, M. Ghavamzadeh, P. Fieguth, A. Khosravi, U. R. Acharya, V. Makarenkov et al., "A review of uncertainty quantification in deep learning: Techniques, applications and challenges," arXiv preprint arXiv:2011.06225, 2020.
  7. I. Biederman, "Recognition-by-components: A theory of human image understanding," Psychological Review, vol. 94, pp. 115-147, 1987.
  8. H. Larochelle, D. Erhan, and Y. Bengio, "Zero-data learning of new tasks." in Proceedings of the Association for the Advancement of Artificial Intelligence, vol. 1, no. 2, 2008, p. 3.
  9. C. H. Lampert, H. Nickisch, and S. Harmeling, "Attribute-based classification for zero-shot visual object categorization," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 36, no. 3, pp. 453-465, 2014.
  10. C. H. Lampert, H. Nickisch, and S. Harmeling, "Learning to detect unseen object classes by between-class attribute transfer," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 951-958.
  11. 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.
  12. Z. Fu, T. Xiang, E. Kodirov, and S. Gong, "Zero-shot learning on semantic class prototype graph," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 8, pp. 2009-2022, 2018.
  13. X. Song, H. Zeng, S. Zhang, L. Herranz, and S. Jiang, "General- ized zero-shot learning with multi-source semantic embeddings for scene recognition," in ACM International Conference on Multi- media, 2020.
  14. H. Zhang, J. Liu, Y. Yao, and Y. Long, "Pseudo distribution on unseen classes for generalized zero shot learning," Pattern Recognition Letters, vol. 135, pp. 451 -458, 2020.
  15. B. Romera-Paredes and P. Torr, "An embarrassingly simple ap- proach to zero-shot learning," in International Conference on Ma- chine Learning, 2015, pp. 2152-2161.
  16. 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 Proceedings of the European Conference on Computer Vision, 2016, pp. 52-68.
  17. W. J. Scheirer, A. de Rezende Rocha, A. Sapkota, and T. E. Boult, "Toward open set recognition," IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, vol. 35, no. 7, pp. 1757-1772, 2012.
  18. Y. Xian, B. Schiele, and Z. Akata, "Zero-shot learning -the good, the bad and the ugly," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 3077-3086.
  19. Y. Xian, C. H. Lampert, B. Schiele, and Z. Akata, "Zero-shot learning-a comprehensive evaluation of the good, the bad and the ugly," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 41, no. 9, pp. 2251-2265, 2018.
  20. K. Liu, W. Liu, H. Ma, W. Huang, and X. Dong, "Generalized zero-shot learning for action recognition with web-scale video data," World Wide Web, vol. 22, no. 2, pp. 807-824, 2019.
  21. Y. Fu, T. Xiang, Y. Jiang, X. Xue, L. Sigal, and S. Gong, "Recent advances in zero-shot recognition: Toward data-efficient under- standing of visual content," IEEE Signal Processing Magazine, vol. 35, no. 1, pp. 112-125, 2018.
  22. M. Rezaei and M. Shahidi, "Zero-shot learning and its applica- tions from autonomous vehicles to covid-19 diagnosis: A review," arXiv preprint arXiv:2004.14143, 2020.
  23. 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.
  24. K. Simonyan and A. Zisserman, "Very deep convolutional networks for large-scale image recognition," arXiv preprint arXiv:1409.1556, 2014.
  25. C. Szegedy, W. Liu, Y. Jia, P. Sermanet, S. Reed, D. Anguelov, D. Erhan, V. Vanhoucke, and A. Rabinovich, "Going deeper with convolutions," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1-9.
  26. P. Zhu, H. Wang, and V. Saligrama, "Generalized zero-shot recog- nition based on visually semantic embedding," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 2995-3003.
  27. H. Zhang, Y. Long, Y. Guan, and L. Shao, "Triple verification network for generalized zero-shot learning," IEEE Transactions on Image Processing, vol. 28, no. 1, pp. 506-517, 2018.
  28. G. Yang, K. Huang, R. Zhang, J. Y. Goulermas, and A. Hussain, "Self-focus deep embedding model for coarse-grained zero-shot classification," in International Conference on Brain Inspired Cogni- tive Systems. Springer, 2019, pp. 12-22.
  29. A. Paul, N. C. Krishnan, and P. Munjal, "Semantically aligned bias reducing zero shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 7056-7065.
  30. A. Cheraghian, S. Rahman, D. Campbell, and L. Petersson, "Transductive zero-shot learning for 3d point cloud classifica- tion," arXiv:1912.07161, 2019.
  31. S. Rahman, S. Khan, and N. Barnes, "Transductive learning for zero-shot object detection," in Proceedings of the IEEE International Conference on Computer Vision, 2019, pp. 6082-6091.
  32. J. Guan, A. Zhao, and Z. Lu, "Extreme reverse projection learning for zero-shot recognition," in Asian Conference on Computer Vision, 2018, pp. 125-141.
  33. Y. Huo, M. Ding, A. Zhao, J. Hu, J.-R. Wen, and Z. Lu, "Zero-shot learning with superclasses," in International Conference on Neural Information Processing. Springer, 2018, pp. 460-472.
  34. T. Long, X. Xu, Y. Li, F. Shen, J. Song, and H. T. Shen, "Pseudo transfer with marginalized corrupted attribute for zero-shot learning," in Proceedings of the ACM International Conference on Multimedia, ser. MM '18, 2018, p. 1802-1810.
  35. L. Zhang, P. Wang, L. Liu, C. Shen, W. Wei, Y. Zhang, and A. van den Hengel, "Towards effective deep embedding for zero- shot learning," IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 9, pp. 2843-2852, 2020.
  36. Y. Xu, X. Xu, G. Han, and S. He, "Holistically-associated trans- ductive zero-shot learning," IEEE Transactions on Cognitive and Developmental Systems, 2021.
  37. T. Mikolov, I. Sutskever, K. Chen, G. Corrado, and J. Dean, "Distributed representations of words and phrases and their compositionality," in Proceedings of the International Conference on Neural Information Processing Systems, ser. NIPS'13, Red Hook, NY, USA, 2013, p. 3111-3119.
  38. E. Kodirov, T. Xiang, and S. Gong, "Semantic autoencoder for zero-shot learning," in Proceedings of the IEEE Conference on Com- puter Vision and Pattern Recognition, 2017, pp. 3174-3183.
  39. F. Wu, S. Zhou, K. Wang, Y. Xu, J. Guan, and J. Huan, "Simple is better: A global semantic consistency based end-to-end frame- work for effective zero-shot learning," in Pacific Rim International Conference on Artificial Intelligence, 2019, pp. 98-112.
  40. Y. Luo, X. Wang, and W. Cao, "A novel dataset-specific feature extractor for zero-shot learning," Neurocomputing, vol. 391, pp. 74 -82, 2020.
  41. F. Wang, J. Liu, S. Zhang, G. Zhang, Y. Li, and F. Yuan, "Inductive zero-shot image annotation via embedding graph," IEEE Access, vol. 7, pp. 107 816-107 830, 2019.
  42. M. Elhoseiny, Y. Zhu, H. Zhang, and A. Elgammal, "Link the head to the" beak": Zero shot learning from noisy text description at part precision," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 6288-6297.
  43. Y. Zhu, M. Elhoseiny, B. Liu, X. Peng, and A. Elgammal, "A gen- erative adversarial approach for zero-shot learning from noisy texts," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 1004-1013.
  44. Y. Le Cacheux, A. Popescu, and H. Le Borgne, "Webly supervised semantic embeddings for large scale zero-shot learning," in Pro- ceedings of the Asian Conference on Computer Vision, 2020.
  45. T. Paz-Argaman, Y. Atzmon, G. Chechik, and R. Tsarfaty, "Zest: Zero-shot learning from text descriptions using textual similarity and visual summarization," arXiv preprint:2010.03276, 2020.
  46. Z. Akata, M. Malinowski, M. Fritz, and B. Schiele, "Multi-cue zero-shot learning with strong supervision," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 59-68.
  47. Y. Liu, X. Gao, Q. Gao, J. Han, and L. Shao, "Label-activating framework for zero-shot learning," Neural Networks, vol. 121, pp. 1-9, 2020.
  48. Y. Shigeto, I. Suzuki, K. Hara, M. Shimbo, and Y. Matsumoto, "Ridge regression, hubness, and zero-shot learning," in Machine Learning and Knowledge Discovery in Databases, 2015, pp. 135-151.
  49. S. Daghaghi, T. Medini, and A. Shrivastava, "Semantic similarity based softmax classifier for zero-shot learning," arXiv preprint arXiv:1909.04790, 2019.
  50. S. Rahman, S. Khan, and F. Porikli, "A unified approach for con- ventional zero-shot, generalized zero-shot, and few-shot learn- ing," IEEE Transactions on Image Processing, vol. 27, no. 11, pp. 5652-5667, 2018.
  51. L. Chen, H. Zhang, J. Xiao, W. Liu, and S.-F. Chang, "Zero- shot visual recognition using semantics-preserving adversarial embedding networks," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 1043-1052.
  52. B. Zhao, X. Sun, Y. Yao, and Y. Wang, "Zero-shot learn- ing via shared-reconstruction-graph pursuit," arXiv preprint arXiv:1711.07302, 2017.
  53. G. Yang, J. Liu, J. Xu, and X. Li, "Dissimilarity representa- tion learning for generalized zero-shot recognition," in Proceed- ings of the ACM International Conference on Multimedia, 2018, p. 2032-2039.
  54. L. Zhang, T. Xiang, and S. Gong, "Learning a deep embedding model for zero-shot learning," in Proceedings of the IEEE Confer- ence on Computer Vision and Pattern Recognition, 2017, pp. 3010- 3019.
  55. Z. Jia, Z. Zhang, L. Wang, C. Shan, and T. Tan, "Deep unbiased embedding transfer for zero-shot learning," IEEE Transactions on Image Processing, vol. 29, pp. 1958-1971, 2020.
  56. V. Kumar Verma, G. Arora, A. Mishra, and P. Rai, "Generalized zero-shot learning via synthesized examples," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 4281-4289.
  57. M. Radovanović, A. Nanopoulos, and M. Ivanović, "Hubs in space: Popular nearest neighbors in high-dimensional data," Journal of Machine Learning Research, vol. 11, no. Sep, pp. 2487- 2531, 2010.
  58. G. Dinu, A. Lazaridou, and M. Baroni, "Improving zero-shot learning by mitigating the hubness problem," arXiv:1412.6568, 2014.
  59. Y. Fu, T. M. Hospedales, T. Xiang, and S. Gong, "Transductive multi-view zero-shot learning," IEEE transactions on Pattern Anal- ysis and Machine Intelligence, vol. 37, no. 11, pp. 2332-2345, 2015.
  60. F. Zhang and G. Shi, "Co-representation network for generalized zero-shot learning," in International Conference on Machine Learn- ing, 2019, pp. 7434-7443.
  61. R. Felix, M. Sasdelli, I. Reid, and G. Carneiro, "Multi-modal ensemble classification for generalized zero shot learning," arXiv preprint arXiv:1901.04623, 2019.
  62. S. Bhattacharjee, D. Mandal, and S. Biswas, "Autoencoder based novelty detection for generalized zero shot learning," in IEEE International Conference on Image Processing, 2019, pp. 3646-3650.
  63. Y. Atzmon and G. Chechik, "Adaptive confidence smoothing for generalized zero-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 11 671-11 680.
  64. S. Min, H. Yao, H. Xie, C. Wang, Z. Zha, and Y. Zhang, "Domain- aware visual bias eliminating for generalized zero-shot learning," in Proceedings of the the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 12 661-12 670.
  65. Y. Le Cacheux, H. Le Borgne, and M. Crucianu, "From classical to generalized zero-shot learning: A simple adaptation process," in International Conference on Multimedia Modeling, 2019, pp. 465-477.
  66. S. Changpinyo, W.-L. Chao, B. Gong, and F. Sha, "Classifier and exemplar synthesis for zero-shot learning," International Journal of Computer Vision, vol. 128, no. 1, pp. 166-201, 2020.
  67. W. Xu, Y. Xian, J. Wang, B. Schiele, and Z. Akata, "Attribute prototype network for zero-shot learning," Advances in Neural Information Processing Systems, vol. 33, 2020.
  68. Y. Guo, G. Ding, J. Han, X. Ding, S. Zhao, Z. Wang, C. Yan, and Q. Dai, "Dual-view ranking with hardness assessment for zero- shot learning," in Proceedings of the Association for the Advancement of Artificial Intelligence, vol. 33, 2019, pp. 8360-8367.
  69. L. Niu, J. Cai, A. Veeraraghavan, and L. Zhang, "Zero-shot learning via category-specific visual-semantic mapping and label refinement," IEEE Transactions on Image Processing, vol. 28, no. 2, pp. 965-979, 2018.
  70. D. Das and C. G. Lee, "Zero-shot image recognition using rela- tional matching, adaptation and calibration," in International Joint Conference on Neural Networks, 2019, pp. 1-8.
  71. D. Huynh and E. Elhamifar, "Fine-grained generalized zero-shot learning via dense attribute-based attention," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 4482-4492.
  72. B. N. Oreshkin, N. Rostamzadeh, P. O. Pinheiro, and C. Pal, "Clarel: Classification via retrieval loss for zero-shot learning," in the IEEE/CVF Conference on Computer Vision and Pattern Recog- nition Workshops, 2020, pp. 3989-3993.
  73. R. Felix, M. Sasdelli, I. Reid, and G. Carneiro, "Augmentation network for generalised zero-shot learning," in Proceedings of the Asian Conference on Computer Vision, 2020.
  74. G. Hinton, O. Vinyals, and J. Dean, "Distilling the knowledge in a neural network," arXiv preprint arXiv:1503.02531, 2015.
  75. X. Wang, S. Pang, and J. Zhu, "Domain segmentation and adjustment for generalized zero-shot learning," arXiv preprint arXiv:2002.00226, 2020.
  76. T. Hayashi and H. Fujita, "Cluster-based zero-shot learning for multivariate data," Journal of ambient intelligence and humanized computing, vol. 12, no. 2, pp. 1897-1911, 2021.
  77. X. Li, M. Ye, L. Zhou, D. Zhang, C. Zhu, and Y. Liu, "Improving generalized zero-shot learning by semantic discriminator," arXiv preprint arXiv:2005.13956, 2020.
  78. R. Felix, B. Harwood, M. Sasdelli, and G. Carneiro, "Generalised zero-shot learning with domain classification in a joint semantic and visual space," in Digital Image Computing: Techniques and Applications, 2019, pp. 1-8.
  79. C. Geng, L. Tao, and S. Chen, "Guided CNN for generalized zero-shot and open-set recognition using visual and semantic prototypes," Pattern Recognition, vol. 102, p. 107263, 2020.
  80. W. J. Scheirer, A. de Rezende Rocha, A. Sapkota, and T. E. Boult, "Toward open set recognition," IEEE Transactions on Pattern Anal- ysis and Machine Intelligence, vol. 35, no. 7, pp. 1757-1772, 2013.
  81. Y. Wang, H. Zhang, Z. Zhang, and Y. Long, "Asymmetric graph based zero shot learning," Multimedia Tools and Applications, pp. 1-22, 2019.
  82. B. Xu, Z. Zeng, C. Lian, and Z. Ding, "Semi-supervised low-rank semantics grouping for zero-shot learning," IEEE Transactions on Image Processing, vol. 30, pp. 2207-2219, 2021.
  83. P. Bhagat, P. Choudhary, and K. M. Singh, "A novel approach based on fully connected weighted bipartite graph for zero-shot learning problems," Journal of Ambient Intelligence and Humanized Computing, pp. 1-16, 2021.
  84. C. Samplawski, E. Learned-Miller, H. Kwon, and B. M. Marlin, "Zero-shot learning in the presence of hierarchically coarsened labels," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020, pp. 926-927.
  85. A. Li, Z. Lu, J. Guan, T. Xiang, L. Wang, and J.-R. Wen, "Trans- ferrable feature and projection learning with class hierarchy for zero-shot learning," International Journal of Computer Vision, vol. 128, pp. 2810-2827, 2020.
  86. X. Wang, Y. Ye, and A. Gupta, "Zero-shot recognition via se- mantic embeddings and knowledge graphs," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 6857-6866.
  87. F. Li, Z. Zhu, X. Zhang, J. Cheng, and Y. Zhao, "From anchor generation to distribution alignment: Learning a discrimina- tive embedding space for zero-shot recognition," arXiv preprint arXiv:2002.03554, 2020.
  88. G.-S. Xie, L. Liu, F. Zhu, F. Zhao, Z. Zhang, Y. Yao, J. Qin, and L. Shao, "Region graph embedding network for zero-shot learning," in European Conference on Computer Vision, 2020, pp. 562-580.
  89. T. N. Kipf and M. Welling, "Semi-supervised classification with graph convolutional networks," arXiv preprint arXiv:1609.02907, 2016.
  90. F. Sung, Y. Yang, L. Zhang, T. Xiang, P. H. S. Torr, and T. M. Hospedales, "Learning to compare: Relation network for few- shot learning," in Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition, 2018, pp. 1199-1208.
  91. R. L. Hu, C. Xiong, and R. Socher, "Correction networks: Meta- learning for zero-shot learning," 2018.
  92. V. K. Verma, K. Liang, N. Mehta, and L. Carin, "Meta-learned at- tribute self-gating for continual generalized zero-shot learning," arXiv preprint:2102.11856, 2021.
  93. V. K. Verma, D. Brahma, and P. Rai, "Meta-learning for general- ized zero-shot learning," in Proceedings of the AAAI Conference on Artificial Intelligence, 2020, pp. 6062-6069.
  94. Y. Yu, Z. Ji, J. Han, and Z. Zhang, "Episode-based prototype generating network for zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 14 035-14 044.
  95. Z. Liu, Y. Li, L. Yao, X. Wang, and G. Long, "Task aligned generative meta-learning for zero-shot learning," arXiv preprint:2103.02185, 2021.
  96. V. K. Verma, A. Mishra, A. Pandey, H. A. Murthy, and P. Rai, "Towards zero-shot learning with fewer seen class examples," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 2241-2251.
  97. G.-S. Xie, L. Liu, X. Jin, F. Zhu, Z. Zhang, J. Qin, Y. Yao, and L. Shao, "Attentive region embedding network for zero-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 9384-9393.
  98. D. Huynh and E. Elhamifar, "A shared multi-attention frame- work for multi-label zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 8776-8786.
  99. Z. Ji, Y. Fu, J. Guo, Y. Pang, Z. M. Zhang et al., "Stacked semantics- guided attention model for fine-grained zero-shot learning," in Advances in Neural Information Processing Systems, 2018, pp. 5995- 6004.
  100. Y. Liu, L. Zhou, X. Bai, Y. Huang, L. Gu, J. Zhou, and T. Harada, "Goal-oriented gaze estimation for zero-shot learning," arXiv preprint arXiv:2103.03433, 2021.
  101. S. Yang, K. Wang, L. Herranz, and J. van de Weijer, "Simple and effective localized attribute representations for zero-shot learning," arXiv preprint arXiv:2006.05938, 2020.
  102. Y. Zhu, J. Xie, Z. Tang, X. Peng, and A. Elgammal, "Semantic- guided multi-attention localization for zero-shot learning," arXiv preprint arXiv:1903.00502, 2019.
  103. Z. Ji, X. Yu, Y. Yu, Y. Pang, and Z. Zhang, "A semantics-guided class imbalance learning model for zero-shot classification," arXiv preprint arXiv:1908.09745, 2019.
  104. M.-C. Yeh and F. Li, "Zero-shot recognition through image- guided semantic classification," arXiv preprint arXiv:2007.11814, 2020.
  105. O. Winn, M. K. Kidambi, and S. Muresan, "Detecting visually relevant sentences for fine-grained classification," in Proceedings of the 5th Workshop on Vision and Language, 2016, pp. 86-91.
  106. I. Misra, A. Gupta, and M. Hebert, "From red wine to red tomato: Composition with context," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 1792-1801.
  107. M. F. Naeem, Y. Xian, F. Tombari, and Z. Akata, "Learning graph embeddings for compositional zero-shot learning," arXiv preprint:2102.01987, 2021.
  108. K. Kato, Y. Li, and A. Gupta, "Compositional learning for human object interaction," in Proceedings of the European Conference on Computer Vision, 2018, pp. 247-264.
  109. X. Chen, A. Shrivastava, and A. Gupta, "Neil: Extracting visual knowledge from web data," in IEEE International Conference on Computer Vision, 2013, pp. 1409-1416.
  110. S. Purushwalkam, M. Nickel, A. Gupta, and M. Ranzato, "Task- driven modular networks for zero-shot compositional learning," Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019.
  111. T. Sylvain, L. Petrini, and D. Hjelm, "Locality and composition- ality in zero-shot learning," in Eighth International Conference on Learning Representations (ICLR), December 2019.
  112. D. Huynh and E. Elhamifar, "Compositional zero-shot learning via fine-grained dense feature composition," in Advances in Neu- ral Information Processing Systems, vol. 33, 2020, pp. 19 849-19 860.
  113. Y. Atzmon, F. Kreuk, U. Shalit, and G. Chechik, "A causal view of compositional zero-shot recognition," in Advances in Neural Information Processing Systems, vol. 33, 2020, pp. 1462-1473.
  114. G. Liu, J. Guan, M. Zhang, J. Zhang, Z. Wang, and Z. Lu, "Joint projection and subspace learning for zero-shot recognition," in IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2019, pp. 1228-1233.
  115. Z. Ji, H. Wang, Y. Pang, and L. Shao, "Dual triplet network for image zero-shot learning," Neurocomputing, vol. 373, pp. 90 -97, 2020.
  116. S. Biswas and Y. Annadani, "Preserving semantic relations for zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 7603-7612.
  117. Y. Yu, Z. Ji, J. Guo, and Z. Zhang, "Zero-shot learning via latent space encoding," IEEE Transactions on Cybernetics, vol. 49, no. 10, pp. 3755-3766, 2018.
  118. J. Li, X. Lan, Y. Liu, L. Wang, and N. Zheng, "Compressing unknown images with product quantizer for efficient zero-shot classification," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 5463-5472.
  119. Y. Liu, Q. Gao, J. Li, J. Han, and L. Shao, "Zero shot learning via low-rank embedded semantic autoencoder," in Proceedings of the International Joint Conference on Artificial Intelligence, 2018, pp. 2490-2496.
  120. Y. Guo, G. Ding, J. Han, and Y. Gao, "SitNet: Discrete similarity transfer network for zero-shot hashing," in Proceedings of the International Joint Conference on Artificial Intelligence, 2017, pp. 1767-1773.
  121. J. Shao and X. Li, "Generalized zero-shot learning with multi- channel gaussian mixture vae," IEEE Signal Processing Letters, vol. 27, pp. 456-460, 2020.
  122. Y. L. Cacheux, H. L. Borgne, and M. Crucianu, "Modeling inter and intra-class relations in the triplet loss for zero-shot learning," in The IEEE International Conference on Computer Vision, 2019.
  123. H. Jiang, R. Wang, S. Shan, and X. Chen, "Transferable contrastive network for generalized zero-shot learning," in Proceedings of the IEEE International Conference on Computer Vision, 2019, pp. 9765- 9774.
  124. X.-B. Jin, G.-S. Xie, K. Huang, H. Cao, and Q.-F. Wang, "Discrim- inant zero-shot learning with center loss," Cognitive Computation, vol. 11, no. 4, pp. 503-512, 2019.
  125. H. Jiang, R. Wang, S. Shan, and X. Chen, "Learning class pro- totypes via structure alignment for zero-shot recognition," in Proceedings of the European Conference on Computer Vision, 2018, pp. 121-138.
  126. X.-B. Jin, G.-S. Xie, K. Huang, J. Miao, and Q. Wang, "Beyond attributes: High-order attribute features for zero-shot learning," in Proceedings of the IEEE International Conference on Computer Vision Workshops, 2019, pp. 0-0.
  127. J. Li, X. Lan, Y. Long, Y. Liu, X. Chen, L. Shao, and N. Zheng, "A joint label space for generalized zero-shot classification," IEEE Transactions on Image Processing, vol. 29, pp. 5817-5831, 2020.
  128. L. Liu, T. Zhou, G. Long, J. Jiang, and C. Zhang, "Attribute prop- agation network for graph zero-shot learning," in Proceedings of the Association for the Advancement of Artificial Intelligence, 2020, pp. 4868-4875.
  129. Y. Wang, H. Zhang, Z. Zhang, Y. Long, and L. Shao, "Learning discriminative domain-invariant prototypes for generalized zero shot learning," Knowledge-Based Systems, vol. 196, p. 105796, 2020.
  130. T. Shen, T. Zhou, G. Long, J. Jiang, S. Pan, and C. Zhang, "Disan: Directional self-attention network for RNN/CNN-free language understanding," in Proceedings of the Association for the Advancement of Artificial Intelligence, 2017.
  131. E. L. Denton, S. Chintala, R. Fergus et al., "Deep generative image models using a laplacian pyramid of adversarial networks," in Advances in Neural Information Processing Systems, 2015, pp. 1486- 1494.
  132. H. Guo and H. L. Viktor, "Learning from imbalanced data sets with boosting and data generation: The databoost-im approach," SIGKDD Explor. Newsl., vol. 6, no. 1, p. 30-39, 2004.
  133. M. Bucher, S. Herbin, and F. Jurie, "Generating visual represen- tations for zero-shot classification," in Proceedings of the IEEE International Conference on Computer Vision Workshops, 2017, pp. 2666-2673.
  134. I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde- Farley, S. Ozair, A. Courville, and Y. Bengio, "Generative adver- sarial nets," in Advances in Neural Information Processing Systems, 2014, pp. 2672-2680.
  135. D. P. Kingma and M. Welling, "Auto-encoding variational bayes," arXiv:1312.6114, 2013.
  136. M. Arjovsky, S. Chintala, and L. Bottou, "Wasserstein generative adversarial networks," in Proceedings of the 34th International Con- ference on Machine Learning, ser. Proceedings of Machine Learning Research, vol. 70, 2017, pp. 214-223.
  137. I. Gulrajani, F. Ahmed, M. Arjovsky, V. Dumoulin, and A. C. Courville, "Improved training of wasserstein GANs," in Advances in Neural Information Processing Systems, 2017, pp. 5767-5777.
  138. Y. Xian, T. Lorenz, B. Schiele, and Z. Akata, "Feature generating networks for zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 5542-5551.
  139. W. Wang, H. Xu, G. Wang, W. Wang, and L. Carin, "Zero-shot recognition via optimal transport," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 3471-3481.
  140. G. Lin, W. Chen, K. Liao, X. Kang, and C. Fan, "Transfer feature generating networks with semantic classes structure for zero-shot learning," IEEE Access, vol. 7, pp. 176 470-176 483, 2019.
  141. C. Finn, P. Abbeel, and S. Levine, "Model-agnostic meta-learning for fast adaptation of deep networks," in International Conference on Machine Learning, 2017, pp. 1126-1135.
  142. J. Li, M. Jing, K. Lu, Z. Ding, L. Zhu, and Z. Huang, "Leveraging the invariant side of generative zero-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 7402-7411.
  143. Y. Ma, X. Xu, F. Shen, and H. T. Shen, "Similarity preserving feature generating networks for zero-shot learning," Neurocom- puting, vol. 406, pp. 333 -342, 2020.
  144. Z. Ye, F. Lyu, L. Li, Q. Fu, J. Ren, and F. Hu, "SR-GAN: Semantic rectifying generative adversarial network for zero-shot learning," in IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2019, pp. 85-90.
  145. A. L. Maas, A. Y. Hannun, and A. Y. Ng, "Rectifier nonlinearities improve neural network acoustic models," in in ICML Workshop on Deep Learning for Audio, Speech and Language Processing, 2013.
  146. M. Elhoseiny and M. Elfeki, "Creativity inspired zero-shot learn- ing," in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), October 2019.
  147. M. Elhoseiny, K. Yi, and M. Elfeki, "Cizsl++: Creativity inspired generative zero-shot learning," arXiv preprint: 2101.00173, 2021.
  148. C. Martindale, The Clockwork Muse: The Predictability of Artistic Change. BasicBooks, 1990.
  149. G.-S. Xie, Z. Zhang, G. Liu, F. Zhu, L. Liu, L. Shao, and X. Li, "Generalized zero-shot learning with multiple graph adaptive generative networks," IEEE Transactions on Neural Networks and Learning Systems, 2021.
  150. H. Xiang, Xie, T. Zeng, and Y. Yang, "Multi-knowledge fusion for new feature generation in generalized zero-shot learning," arXiv preprint arXiv:2102.11566, 2021.
  151. C. Xie, H. Xiang, T. Zeng, Y. Yang, B. Yu, and Q. Liu, "Cross knowledge-based generative zero-shot learning approach with taxonomy regularization," Neural Networks, vol. 139, pp. 168-178, 2021.
  152. Y. Li, K. Swersky, and R. Zemel, "Generative moment matching networks," in International Conference on Machine Learning, 2015, pp. 1718-1727.
  153. A. Odena, C. Olah, and J. Shlens, "Conditional image synthesis with auxiliary classifier GANs," in Proceedings of the 34th Interna- tional Conference on Machine Learning, vol. 70. JMLR. org, 2017, pp. 2642-2651.
  154. Y. Bengio, L. Yao, G. Alain, and P. Vincent, "Generalized denois- ing auto-encoders as generative models," in Advances in Neural Information Processing Systems, 2013, pp. 899-907.
  155. A. Makhzani, J. Shlens, N. Jaitly, I. Goodfellow, and B. Frey, "Adversarial autoencoders," arXiv:1511.05644, 2015.
  156. J. Ni, S. Zhang, and H. Xie, "Dual adversarial semantics- consistent network for generalized zero-shot learning," CoRR, vol. abs/1907.05570, 2019.
  157. J.-Y. Zhu, T. Park, P. Isola, and A. A. Efros, "Unpaired image-to- image translation using cycle-consistent adversarial networks," in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 2223-2232.
  158. Z. Chen, J. Li, Y. Luo, Z. Huang, and Y. Yang, "Canzsl: Cycle- consistent adversarial networks for zero-shot learning from nat- ural language," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2020, pp. 874-883.
  159. R. Felix, V. B. Kumar, I. Reid, and G. Carneiro, "Multi-modal cycle-consistent generalized zero-shot learning," in Proceedings of the European Conference on Computer Vision, 2018, pp. 21-37.
  160. J. Li, M. Jing, K. Lu, L. Zhu, Y. Yang, and Z. Huang, "Alleviating feature confusion for generative zero-shot learning," in Proceed- ings of the 27th ACM International Conference on Multimedia, 2019, pp. 1587-1595.
  161. J. Li, M. Jing, K. Lu, L. Zhu, and H. T. Shen, "Investigating the bilateral connections in generative zero-shot learning," IEEE Transactions on Cybernetics, pp. 1-12, 2021.
  162. S. Chandhok and V. N. Balasubramanian, "Two-level adversarial visual-semantic coupling for generalized zero-shot learning," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 3100-3108.
  163. K. Sohn, H. Lee, and X. Yan, "Learning structured output repre- sentation using deep conditional generative models," in Advances in Neural Information Processing Systems, 2015, pp. 3483-3491.
  164. A. Mishra, S. Krishna Reddy, A. Mittal, and H. A. Murthy, "A generative model for zero shot learning using conditional variational autoencoders," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2018, pp. 2188-2196.
  165. P. Zhu, H. Wang, and V. Saligrama, "Don't even look once: Syn- thesizing features for zero-shot detection," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 11 693-11 702.
  166. R. Gao, X. Hou, J. Qin, J. Chen, L. Liu, F. Zhu, Z. Zhang, and L. Shao, "Zero-VAE-GAN: Generating unseen features for gener- alized and transductive zero-shot learning," IEEE Transactions on Image Processing, vol. 29, pp. 3665-3680, 2020.
  167. H. Huang, C. Wang, P. S. Yu, and C.-D. Wang, "Generative dual adversarial network for generalized zero-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 801-810.
  168. 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.
  169. Z. Chen, Z. Huang, J. Li, and Z. Zhang, "Entropy-based un- certainty calibration for generalized zero-shot learning," arXiv preprint:2101.03292, 2021.
  170. C. Li, X. Ye, H. Yang, Y. Han, X. Li, and Y. Jia, "Generalized zero shot learning via synthesis pseudo features," IEEE Access, vol. 7, pp. 87 827-87 836, 2019.
  171. A. B. L. Larsen, S. K. Sønderby, H. Larochelle, and O. Winther, "Autoencoding beyond pixels using a learned similarity metric," in International conference on machine learning. PMLR, 2016, pp. 1558-1566.
  172. T. Salimans, I. Goodfellow, W. Zaremba, V. Cheung, A. Radford, and X. Chen, "Improved techniques for training gans," in Proceed- ings of the International Conference on Neural Information Processing Systems, 2016, p. 2234-2242.
  173. R. Gao, X. Hou, J. Qin, L. Liu, F. Zhu, and Z. Zhang, "A joint generative model for zero-shot learning," in Proceedings of the European Conference on Computer Vision, 2018, pp. 0-0.
  174. J. Johnson, A. Alahi, and L. Fei-Fei, "Perceptual losses for real- time style transfer and super-resolution," in Proceedings of the European Conference on Computer Vision. Springer, 2016, pp. 694- 711.
  175. T. Xu, Y. Zhao, and X. Liu, "Dual generative network with discriminative information for generalized zero-shot learning," Complexity, vol. 2021, 2021.
  176. C. Tang, Y. Kuang, J. Lv, and J. Hu, "San: Sampling adversarial networks for zero-shot learning," in International Conference on Neural Information Processing. Springer, 2020, pp. 626-638.
  177. X. Li, M. Fang, H. Li, and J. Wu, "Learning domain invariant un- seen features for generalized zero-shot classification," Knowledge- Based Systems, vol. 206, p. 106378, 2020.
  178. A. K. Pambala, T. Dutta, and S. Biswas, "Generative model with semantic embedding and integrated classifier for generalized zero-shot learning," in IEEE Winter Conference on Applications of Computer Vision, 2020, pp. 1226-1235.
  179. Y. Yu, Z. Ji, Y. Pang, J. Guo, Z. Zhang, and F. Wu, "Bi- adversarial auto-encoder for zero-shot learning," arXiv preprint arXiv:1811.08103, 2018.
  180. Z. Ji, G. Dai, and Y. Yu, "Multi-modality adversarial auto-encoder for zero-shot learning," IEEE Access, vol. 8, pp. 9287-9295, 2019.
  181. X. Luo, Z. Cai, F. Wu, and J. Xiao-Yuan, "A generative model for zero-shot learning via wasserstein auto-encoder," in IEEE International Conference on Information Technology, Big Data and Artificial Intelligence, vol. 1, 2020, pp. 757-762.
  182. L. Feng and C. Zhao, "Transfer increment for generalized zero- shot learning," IEEE Transactions on Neural Networks and Learning Systems, pp. 1-15, 2020.
  183. Y. Long, L. Liu, F. Shen, L. Shao, and X. Li, "Zero-shot learning using synthesised unseen visual data with diffusion regularisa- tion," IEEE Transactions on Pattern Analysis and Machine Intelli- gence, vol. 40, no. 10, pp. 2498-2512, 2018.
  184. Z. Ji, J. Wang, Y. Yu, Y. Pang, and J. Han, "Class-specific synthe- sized dictionary model for zero-shot learning," Neurocomputing, vol. 329, pp. 339 -347, 2019.
  185. Y. Shi and W. Wei, "Discriminative embedding autoencoder with a regressor feedback for zero-shot learning," IEEE Access, vol. 8, pp. 11 019-11 030, 2020.
  186. S. Changpinyo, W.-L. Chao, and F. Sha, "Predicting visual exem- plars of unseen classes for zero-shot learning," in Proceedings of the IEEE international conference on computer vision, 2017, pp. 3476- 3485.
  187. J. Song, C. Shen, Y. Yang, Y. Liu, and M. Song, "Transductive unbiased embedding for zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 1024-1033.
  188. Y. Li, D. Wang, H. Hu, Y. Lin, and Y. Zhuang, "Zero-shot recog- nition using dual visual-semantic mapping paths," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 3279-3287.
  189. H. Zhang, L. Liu, Y. Long, Z. Zhang, and L. Shao, "Deep transductive network for generalized zero shot learning," Pattern Recognition, vol. 105, p. 107370, 2020.
  190. X. Xu, F. Shen, Y. Yang, D. Zhang, H. Tao Shen, and J. Song, "Matrix tri-factorization with manifold regularizations for zero- shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 3798-3807.
  191. C. Ding, T. Li, W. Peng, and H. Park, "Orthogonal nonnegative matrix t-factorizations for clustering," in Proceedings of the Inter- national Conference on Knowledge Discovery and Data Mining, 2006, pp. 126-135.
  192. Y. Xian, S. Sharma, B. Schiele, and Z. Akata, "f-vaegan-d2: A fea- ture generating framework for any-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 10 275-10 284.
  193. J. Wu, T. Zhang, Z.-J. Zha, J. Luo, Y. Zhang, and F. Wu, "Self- supervised domain-aware generative network for generalized zero-shot learning," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 12 767-12 776.
  194. G. A. Miller, "WordNet: a lexical database for english," Commu- nications of the ACM, vol. 38, no. 11, pp. 39-41, 1995.
  195. M. Norouzi, T. Mikolov, S. Bengio, Y. Singer, J. Shlens, A. Frome, G. S. Corrado, and J. Dean, "Zero-shot learning by convex combi- nation of semantic embeddings," arXiv preprint arXiv:1312.5650, 2013.
  196. S. Changpinyo, W.-L. Chao, B. Gong, and F. Sha, "Synthesized classifiers for zero-shot learning," in Proceedings of the IEEE con- ference on computer vision and pattern recognition, 2016, pp. 5327- 5336.
  197. C. Wah, S. Branson, P. Welinder, P. Perona, and S. Belongie, "The Caltech-UCSD Birds-200-2011 Dataset," California Institute of Technology, Tech. Rep. CNS-TR-2011-001, 2011.
  198. P. Welinder, S. Branson, T. Mita, C. Wah, F. Schroff, S. Belongie, and P. Perona, "Caltech-ucsd birds 200," California Institute of Technology, Tech. Rep., 2010.
  199. A. Farhadi, I. Endres, D. Hoiem, and D. Forsyth, "Describing objects by their attributes," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 1778-1785.
  200. C. H. Lampert, H. Nickisch, and S. Harmeling, "Attribute-based classification for zero-shot visual object categorization," IEEE transactions on pattern analysis and machine intelligence, vol. 36, no. 3, pp. 453-465, 2013.
  201. J. Xiao, J. Hays, K. A. Ehinger, A. Oliva, and A. Torralba, "Sun database: Large-scale scene recognition from abbey to zoo," in IEEE computer society conference on computer vision and pattern recognition, 2010, pp. 3485-3492.
  202. G. Patterson and J. Hays, "Sun attribute database: Discovering, annotating, and recognizing scene attributes," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2012, pp. 2751-2758.
  203. G. Patterson, C. Xu, H. Su, and J. Hays, "The sun attribute database: Beyond categories for deeper scene understanding," International Journal of Computer Vision, vol. 108, no. 1-2, pp. 59- 81, 2014.
  204. G. Van Horn, S. Branson, R. Farrell, S. Haber, J. Barry, P. Ipeirotis, P. Perona, and S. Belongie, "Building a bird recognition app and large scale dataset with citizen scientists: The fine print in fine- grained dataset collection," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015, pp. 595-604.
  205. Z. Liu, P. Luo, S. Qiu, X. Wang, and X. Tang, "Deepfashion: Powering robust clothes recognition and retrieval with rich an- notations," in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 1096-1104.
  206. B. Zhao, Y. Fu, R. Liang, J. Wu, Y. Wang, and Y. Wang, "A large- scale attribute dataset for zero-shot learning," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops, 2019, pp. 0-0.
  207. A. Khosla, N. Jayadevaprakash, B. Yao, and F.-F. Li, "Novel dataset for fine-grained image categorization: Stanford dogs," in Proceedings of the IEEE conference on computer vision and pattern recognition, vol. 2, no. 1, 2011.
  208. M.-E. Nilsback and A. Zisserman, "Automated flower classifi- cation over a large number of classes," in Indian Conference on Computer Vision, Graphics & Image Processing, 2008, pp. 722-729.
  209. P. Zhu, H. Wang, and V. Saligrama, "Don't even look once: Syn- thesizing features for zero-shot detection," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 11 690-11 699.
  210. M. Vyas, "Zero shot learning for visual object recognition with generative models," Ph.D. dissertation, Arizona State University, 2020.
  211. K. Lee, K. Lee, K. Min, Y. Zhang, J. Shin, and H. Lee, "Hierarchical novelty detection for visual object recognition," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018, pp. 1034-1042.
  212. E. Zablocki, P. Bordes, L. Soulier, B. Piwowarski, and P. Gallinari, "Context-aware zero-shot learning for object recognition," in Proceedings of Machine Learning Research, vol. 97, 2019, pp. 7292- 7303.
  213. A. Cheraghian, S. Rahman, D. Campbell, and L. Petersson, "Miti- gating the hubness problem for zero-shot learning of 3d objects," arXiv preprint arXiv:1907.06371, 2019.
  214. Y. L. Cacheux, H. L. Borgne, and M. Crucianu, "Zero-shot learn- ing with deep neural networks for object recognition," arXiv preprint arXiv:2102.03137, 2021.
  215. C. Lee, W. Fang, C. Yeh, and Y. F. Wang, "Multi-label zero-shot learning with structured knowledge graphs," in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 1576-1585.
  216. A. Gupta, S. Narayan, S. Khan, F. S. Khan, L. Shao, and J. van de Weijer, "Generative multi-label zero-shot learning," arXiv preprint:2101.11606, 2021.
  217. A. Paul, T. C. Shen, S. Lee, N. Balachandar, Y. Peng, Z. Lu, and R. M. Summers, "Generalized zero-shot chest x-ray diagnosis through trait-guided multi-view semantic embedding with self- training," IEEE Transactions on Medical Imaging, pp. 1-1, 2021.
  218. G. Wen, J. Ma, Y. Hu, H. Li, and L. Jiang, "Grouping attributes zero-shot learning for tongue constitution recognition," Artificial Intelligence in Medicine, vol. 109, p. 101951, 2020.
  219. M. Bucher, V. Tuan-Hung, M. Cord, and P. Pérez, "Zero-shot semantic segmentation," in Advances in Neural Information Pro- cessing Systems, 2019, pp. 466-477.
  220. Y. Zhang and O. Khriyenko, "Zero-shot semantic segmentation using relation network," in Conference of Open Innovations Associ- ation, 2021, pp. 516-527.
  221. T. Dutta and S. Biswas, "Generalized zero-shot cross-modal re- trieval," IEEE Transactions on Image Processing, vol. 28, no. 12, pp. 5953-5962, 2019.
  222. J. Zhu, X. Xu, F. Shen, R. K. Lee, Z. Wang, and H. T. Shen, "OCEAN: A dual learning approach for generalized zero-shot sketch-based image retrieval," in IEEE International Conference on Multimedia and Expo, 2020, pp. 1-6.
  223. D. Mandal, S. Narayan, S. K. Dwivedi, V. Gupta, S. Ahmed, F. S. Khan, and L. Shao, "Out-of-distribution detection for generalized zero-shot action recognition," in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019, pp. 9985-9993.
  224. S. N. Gowda, L. Sevilla-Lara, F. Keller, and M. Rohrbach, "CLASTER: Clustering with reinforcement learning for zero-shot action recognition," arXiv preprint:2101.07042, 2021.
  225. Q. Wang and K. Chen, "Multi-label zero-shot human action recognition via joint latent ranking embedding," Neural Networks, vol. 122, pp. 1-23, 2020.
  226. K. Parida, N. Matiyali, T. Guha, and G. Sharma, "Coordinated joint multimodal embeddings for generalized audio-visual zero- shot classification and retrieval of videos," in The IEEE Winter Conference on Applications of Computer Vision, 2020, pp. 3251-3260.
  227. P. Mazumder, P. Singh, K. K. Parida, and V. P. Namboodiri, "AVGZSLNet: audio-visual generalized zero-shot learning by reconstructing label features from multi-modal embeddings," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 3090-3099.
  228. M. E. Basiri, M. Abdar, M. A. Cifci, S. Nemati, and U. R. Acharya, "A novel method for sentiment classification of drug reviews using fusion of deep and machine learning techniques," Knowledge-Based Systems, p. 105949, 2020.
  229. M. E. Basiri, S. Nemati, M. Abdar, E. Cambria, and U. R. Achar- rya, "ABCDM: An attention-based bidirectional CNN-RNN deep model for sentiment analysis," Future Generation Computer Sys- tems, 2020.
  230. M. Abdar, M. E. Basiri, J. Yin, M. Habibnezhad, G. Chi, S. Ne- mati, and S. Asadi, "Energy choices in alaska: Mining people's perception and attitudes from geotagged tweets," Renewable and Sustainable Energy Reviews, vol. 124, p. 109781, 2020.
  231. C. Song, S. Zhang, N. Sadoughi, P. Xie, and E. Xing, "Generalized zero-shot text classification for icd coding," in Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, 2020, pp. 4018-4024.
  232. L. K. Hansen and P. Salamon, "Neural network ensembles," IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 12, no. 10, pp. 993-1001, 1990.
  233. F. Pourpanah, R. Wang, C. P. Lim, X. Wang, M. Seera, and C. J. Tan, "An improved fuzzy artmap and q-learning agent model for pattern classification," Neurocomputing, vol. 359, pp. 139 -152, 2019.
  234. F. Pourpanah, C. J. Tan, C. P. Lim, and J. Mohamad-Saleh, "A q-learning-based multi-agent system for data classification," Applied Soft Computing, vol. 52, pp. 519 -531, 2017.
  235. R. S. Sutton and A. G. Barto, Introduction to reinforcement learning. MIT press Cambridge, 1998, vol. 135.
  236. F. Pourpanah, C. P. Lim, and Q. Hao, "A reinforced fuzzy artmap model for data classification," International Journal of Machine Learning Cybernetics, vol. 10, no. 7, pp. 1643-1655, 2019.
  237. D. Samuel, Y. Atzmon, and G. Chechik, "From generalized zero- shot learning to long-tail with class descriptors," in Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2021, pp. 286-295.

Related papers

A Unified approach for Conventional Zero-shot, Generalized Zero-shot and Few-shot Learning
Shafin Rahman

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society, 2018

Prevalent techniques in zero-shot learning do not generalize well to other related problem scenarios. Here, we present a unified approach for conventional zero-shot, generalized zero-shot and few-shot learning problems. Our approach is based on a novel Class Adapting Principal Directions (CAPD) concept that allows multiple embeddings of image features into a semantic space. Given an image, our method produces one principal direction for each seen class. Then, it learns how to combine these directions to obtain the principal direction for each unseen class such that the CAPD of the test image is aligned with the semantic embedding of the true class, and opposite to the other classes. This allows efficient and class-adaptive information transfer from seen to unseen classes. In addition, we propose an automatic process for selection of the most useful seen classes for each unseen class to achieve robustness in zero-shot learning. Our method can update the unseen CAPD taking the advanta...

View PDFchevron_right
Generalized Zero-Shot Learning using Generated Proxy Unseen Samples and Entropy Separation
Fabio Cuzzolin

Proceedings of the 28th ACM International Conference on Multimedia

The recent generative model-driven Generalized Zero-shot Learning (GZSL) techniques overcome the prevailing issue of the model bias towards the seen classes by synthesizing the visual samples of the unseen classes through leveraging the corresponding semantic prototypes. Although such approaches significantly improve the GZSL performance due to data augmentation, they violate the principal assumption of GZSL regarding the unavailability of semantic information of unseen classes during training. In this work, we propose to use a generative model (GAN) for synthesizing the visual proxy samples while strictly adhering to the standard assumptions of the GZSL. The aforementioned proxy samples are generated by exploring the early training regime of the GAN. We hypothesize that such proxy samples can effectively be used to characterize the average entropy of the label distribution of the samples from the unseen classes. Further, we train a classifier on the visual samples from the seen classes and proxy samples using entropy separation criterion such that an average entropy of the label distribution is low and high, respectively, for the visual samples from the seen classes and the proxy samples. Such entropy separation criterion generalizes well during testing where the samples from the unseen classes exhibit higher entropy than the entropy of the samples from the seen classes. Subsequently, low and high entropy samples are classified using supervised learning and ZSL rather than GZSL. We show the superiority of the proposed method by experimenting on AWA1, CUB, HMDB51, and UCF101 datasets. CCS CONCEPTS • Computing methodologies → Object recognition.

View PDFchevron_right
SDM-Net: A Simple and Effective Model for Generalized Zero-Shot Learning
Tharun Medini

2021

Zero-Shot Learning (ZSL) is a classification task where some classes referred to as unseen classes have no training images. Instead, we only have side information about seen and unseen classes, often in the form of semantic or descriptive attributes. Lack of training images from a set of classes restricts the use of standard classification techniques and losses, including the widespread cross-entropy loss. We introduce a novel Similarity Distribution Matching Network (SDM-Net) which is a standard fully connected neural network architecture with a non-trainable penultimate layer consisting of class attributes. The output layer of SDM-Net consists of both seen and unseen classes. To enable zero-shot learning, during training, we regularize the model such that the predicted distribution of unseen class is close in KL divergence to the distribution of similarities between the correct seen class and all the unseen classes. We evaluate the proposed model on five benchmark datasets for zer...

View PDFchevron_right
Zero-shot Learning for Visual Recognition Problems
Shujon Naha

2015

In this thesis we discuss different aspects of zero-shot learning and propose solutions for three challenging visual recognition problems: 1) unknown object recognition from images 2) novel action recognition from videos and 3) unseen object segmentation. In all of these three problems, we have two different sets of classes, the “known classes”, which are used in the training phase and the “unknown classes” for which there is no training instance. Our proposed approach exploits the available semantic relationships between known and unknown object classes and use them to transfer the appearance models from known object classes to unknown object classes to recognize unknown objects. We also propose an approach to recognize novel actions from videos by learning a joint model that links videos and text. Finally, we present a ranking based approach for zero-shot object segmentation. We represent each unknown object class as a semantic ranking of all the known classes and use this semanti...

View PDFchevron_right
Trading-off Information Modalities in Zero-shot Classification
Matias Molina

2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022

Zero-shot classification is the task of learning predictors for classes not seen during training. A practical way to deal with the lack of annotations for the target categories is to encode not only the inputs (images) but also the outputs (object classes) into a suitable representation space. We can use these representations to measure the degree at which images and categories agree by fitting a compatibility measure using the information available during training. One way to define such a measure is by a two step process in which we first project the elements of either space (visual or semantic) onto the other and then compute a similarity score in the target space. Although projections onto the visual space has shown better general performance, little attention has been paid to the degree at which the visual and semantic information contribute to the final predictions. In this paper, we build on this observation and propose two different formulations that allow us to explicitly trade-off the relative importance of the visual and semantic spaces for classification in a zero-shot setting. Our formulations are based on redefinition of the similarity scoring and loss function used to learn the projections. Experiments on six different datasets show that our approach lead to improve performance compared to similar methods. Moreover, combined with synthetic features, our approach competes favorably with the state of the art on both the standard and generalized settings.

View PDFchevron_right
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved