Papers › DUET: Cross-modal Semantic Grounding for Contrastive Zero-shot Learning

DUET: Cross-modal Semantic Grounding for Contrastive Zero-shot Learning

4 Jul 2022arXiv:2207.01328archive 2025-07-28

Zhuo Chen, Yufeng Huang, Jiaoyan Chen, Yuxia Geng, Wen Zhang, Yin Fang, Jeff Z. Pan, Huajun Chen

Zero-shot learning (ZSL) aims to predict unseen classes whose samples have never appeared during training. One of the most effective and widely used semantic information for zero-shot image classification are attributes which are annotations for class-level visual characteristics. However, the current methods often fail to discriminate those subtle visual distinctions between images due to not only the shortage of fine-grained annotations, but also the attribute imbalance and co-occurrence. In this paper, we present a transformer-based end-to-end ZSL method named DUET, which integrates latent semantic knowledge from the pre-trained language models (PLMs) via a self-supervised multi-modal learning paradigm. Specifically, we (1) developed a cross-modal semantic grounding network to investigate the model's capability of disentangling semantic attributes from the images; (2) applied an attribute-level contrastive learning strategy to further enhance the model's discrimination on fine-grained visual characteristics against the attribute co-occurrence and imbalance; (3) proposed a multi-task learning policy for considering multi-model objectives. We find that our DUET can achieve state-of-the-art performance on three standard ZSL benchmarks and a knowledge graph equipped ZSL benchmark. Its components are effective and its predictions are interpretable.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

zjukg/DUET officialmentioned in papermentioned on GitHubpytorchMIT report
zjukg/structure-clip mentioned on GitHubpytorch report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

AttributeContrastive LearningImage ClassificationMulti-Task LearningZero-Shot Image ClassificationZero-Shot Learningimage-classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Zero-Shot Learning AwA2 DUET (Ours) Accuracy Seen 84.7 #4 of 4 Archive leaderboard report
Zero-Shot Learning AwA2 DUET (Ours) Accuracy Unseen 63.7 #4 of 4 Archive leaderboard report
Zero-Shot Learning AwA2 DUET (Ours) H 72.7 #4 of 4 Archive leaderboard report
Zero-Shot Learning AwA2 DUET (Ours) average top-1 classification accuracy 69.9 #4 of 4 Archive leaderboard report
Zero-Shot Learning CUB-200-2011 DUET Accuracy Seen 72.8 #2 of 14 Archive leaderboard report
Zero-Shot Learning CUB-200-2011 DUET Accuracy Unseen 62.9 #2 of 14 Archive leaderboard report
Zero-Shot Learning CUB-200-2011 DUET H 67.5 #2 of 14 Archive leaderboard report
Zero-Shot Learning CUB-200-2011 DUET average top-1 classification accuracy 72.3 #2 of 14 Archive leaderboard report
Zero-Shot Learning SUN Attribute DUET (Ours) Accuracy Seen 45.8 #5 of 9 Archive leaderboard report
Zero-Shot Learning SUN Attribute DUET (Ours) Accuracy Unseen 45.7 #5 of 9 Archive leaderboard report
Zero-Shot Learning SUN Attribute DUET (Ours) H 45.8 #5 of 9 Archive leaderboard report
Zero-Shot Learning SUN Attribute DUET (Ours) average top-1 classification accuracy 64.4 #5 of 9 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

Concatenated Skip ConnectionContrastive LearningSoftmax

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections