GNDAN: Graph Navigated Dual Attention Network for Zero-Shot Learning

Chen, Shiming; Hong, Ziming; Xie, Guo-Sen; Peng, Qinmu; You, Xinge; Ding, Weiping; Shao, Ling

doi:10.1109/tnnls.2022.3155602

Cited by 25 publications

(8 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Baselines. We choose f-CLSWGAN (Xian et al 2018b), Cycle-WGAN (Felix et al 2018), LisGAN (Li et al 2019), TCN (Jiang et al 2019b), f-VAEGAN (Xian et al 2019b), TF-VAEGAN (Narayan et al 2020), GCMCF (Yue et al 2021), HSVA (Chen et al 2021a) MSDN (Chen et al 2022b) AZSL (Gao et al 2022) and SHIP+CoOp (Wang et al 2023) as our baseline methods.…”

Section: Results Of Generalized Zero-shot Learningmentioning

confidence: 99%

Data-Free Generalized Zero-Shot Learning

Tang,

Zhang,

Yan

et al. 2024

AAAI

View full text Add to dashboard Cite

Deep learning models have the ability to extract rich knowledge from large-scale datasets. However, the sharing of data has become increasingly challenging due to concerns regarding data copyright and privacy. Consequently, this hampers the effective transfer of knowledge from existing data to novel downstream tasks and concepts. Zero-shot learning (ZSL) approaches aim to recognize new classes by transferring semantic knowledge learned from base classes. However, traditional generative ZSL methods often require access to real images from base classes and rely on manually annotated attributes, which presents challenges in terms of data restrictions and model scalability. To this end, this paper tackles a challenging and practical problem dubbed as data-free zero-shot learning (DFZSL), where only the CLIP-based base classes data pre-trained classifier is available for zero-shot classification. Specifically, we propose a generic framework for DFZSL, which consists of three main components. Firstly, to recover the virtual features of the base data, we model the CLIP features of base class images as samples from a von Mises-Fisher (vMF) distribution based on the pre-trained classifier. Secondly, we leverage the text features of CLIP as low-cost semantic information and propose a feature-language prompt tuning (FLPT) method to further align the virtual image features and textual features. Thirdly, we train a conditional generative model using the well-aligned virtual image features and corresponding semantic text features, enabling the generation of new classes features and achieve better zero-shot generalization. Our framework has been evaluated on five commonly used benchmarks for generalized ZSL, as well as 11 benchmarks for the base-to-new ZSL. The results demonstrate the superiority and effectiveness of our approach. Our code is available in https://github.com/ylong4/DFZSL.

show abstract

Section: Results Of Generalized Zero-shot Learningmentioning

confidence: 99%

Data-Free Generalized Zero-Shot Learning

Tang,

Zhang,

Yan

et al. 2024

AAAI

View full text Add to dashboard Cite

show abstract

“…This limitation makes single-label zero-shot learning difficult to generalize to multi-label scenarios, especially when multiple unseen labels need to be predicted. Chen et al [50] proposed a graph-navigated dual attention network to jointly learn local and explicit global embeddings with a region-guided attention network and region-guided graph attention network. A selfcalibration mechanism is designed to improve the visualsemantic interaction and prevent the overfitting of unseen classes.…”

Section: B Zero-shot Learningmentioning

confidence: 99%

Towards Unbiased Multi-Label Zero-Shot Learning With Pyramid and Semantic Attention

Liu

Guo

et al. 2023

IEEE Trans. Multimedia

View full text Add to dashboard Cite

This paper investigates a challenging problem of zero-shot learning in the multi-label scenario (MLZSL), wherein, the model is trained to recognize multiple unseen classes within a sample (e.g., an image) based on seen classes and auxiliary knowledge, e.g., semantic information. Existing methods usually resort to analyzing the relationship of various seen classes residing in a sample from the dimension of spatial or semantic characteristics, and transfer the learned model to unseen ones. But they ignore the effective integration of local and global features. That is, in the process of inferring unseen classes, global features represent the principal direction of the image in the feature space, while local features should maintain uniqueness within a certain range. This integrated neglect will make the model lose its grasp of the main components of the image. Relying only on the local existence of seen classes during the inference stage introduces unavoidable bias. In this paper, we propose a novel and effective group bienhancement framework for MLZSL, dubbed GBE-MLZSL, to fully make use of such properties and enable a more accurate and robust visual-semantic projection. Specifically, we split the feature maps into several feature groups, of which each feature group can be trained independently with the Local Information Distinguishing Module (LID) to ensure uniqueness. Meanwhile, a Global Enhancement Module (GEM) is designed to preserve the principal direction. Besides, a static graph structure is designed to construct the correlation of local features. Experiments on largescale MLZSL benchmark datasets NUS-WIDE and Open-Images-v4 demonstrate that the proposed GBE-MLZSL outperforms other state-of-the-art methods with large margins.

show abstract

“…And the most interesting application for us is that GNN could perform well in the zero-shot learning (ZSL). (Gao and Xu, 2020) (Chen et al, 2022) And in zero-shot task, GNN can help model relationships such as relationships between text descriptions well. And there are still other domains that GNN could be applied to such as traffic control , logic reasoning (Zhang et al, 2019), adversarial attack prevention , social influence prediction (Song et al, 2021), etc.…”

Section: Applications Of Gnnmentioning

confidence: 99%

A Survey of GNN in Bioinformation Data

Zhu¹

2022

Preprint

View full text Add to dashboard Cite

With the development of data science, more and more machine learning technologies have been designed to solve complicated and challenging real-world tasks containing a large volume of data. And many significant real-world datasets contain data in the form of networks or graphs. Graph Neural Networks is one of the powerful machine learning tools, which could provide a perfect solution to processing a large amount of non-Euclidean data. And because most bio information data in bioinformatics is in the non-Euclidean domain, Graph Neural Networks could then directly be applied to solve problems in bioinformatics. Much research has been done in the field of GNN, and there are also some surveys related to GNN and its applications. However, there has been little research focusing on GNN in bioinformatics, and we think in the future we could better utilize GNN in the field of biology, so we would like to write a literature review to help take a comprehensive look at GNN and their applications in the field of bioinformatics. In this paper, we would first introduce SOTA models in Graph Neural Networks, and second, introduce their applications in bio information. And then we would provide future directions of Graph Neural Networks in bioinformatics.

show abstract

GNDAN: Graph Navigated Dual Attention Network for Zero-Shot Learning

Cited by 25 publications

References 63 publications

Data-Free Generalized Zero-Shot Learning

Data-Free Generalized Zero-Shot Learning

Towards Unbiased Multi-Label Zero-Shot Learning With Pyramid and Semantic Attention

A Survey of GNN in Bioinformation Data

Contact Info

Product

Resources

About