Bi-Directional Distribution Alignment for Transductive Zero-Shot Learning

Zhicai Wang, Y. Hao, Tingting Mu, Ouxiang Li, Shuo Wang +1 more
2/14/2026

No official code found. Best implementation: Zhicaiwww/Bi-VAEGAN (35 stars, last commit 2024-04-10).

Abstract

Zero-shot learning (ZSL) suffers intensely from the domain shift issue, i.e., the mismatch (or misalignment) between the true and learned data distributions for classes without training data (unseen classes). By learning additionally from unlabelled data collected for the unseen classes, transductive ZSL (TZSL) could reduce the shift but only to a certain extent. To improve TZSL, we propose a novel approach Bi-VAEGAN which strengthens the distribution alignment between the visual space and an auxiliary space. As a result, it can reduce largely the domain shift. The proposed key designs include (1) a bi-directional distribution alignment, (2) a simple but effective L2-norm based feature normalization approach, and (3) a more sophisticated unseen class prior estimation. Evaluated by four benchmark datasets, Bi-VAEGAN11Code is available at https://github.com/Zhicaiwww/Bi-VAEGAN achieves the new state of the art under both the standard and generalized TZSL settings.

View on arXivView PDF

Code Implementations(1)

Zhicaiwww/Bi-VAEGANCommunity100%

Official implementation of CVPR2023 paper "Bi-directional distribution alignment for transductive zero-shot learning""

357Shell, PythonMar 7, 20232 years ago

Similar Papers

Domain Adaptive Few-Shot Open-Set Learning

86% match1 repo172026

Enabling Generalized Zero-shot Learning Towards Unseen Domains by Intrinsic Learning from Redundant LLM Semantics

85% match2024

Step-wise Distribution Alignment Guided Style Prompt Tuning for Source-free Cross-domain Few-shot Learning

82% match1 repo152024

Extremely Simple Out-of-distribution Detection for Audio-visual Generalized Zero-shot Learning

82% match2025

Improved Feature Generating Framework for Transductive Zero-shot Learning

82% match2024

Covariance-based Space Regularization for Few-shot Class Incremental Learning

81% match1 repo52024

Cite this paper

@article{wang2026bidirectional,
  title  = {Bi-Directional Distribution Alignment for Transductive Zero-Shot Learning},
  author = {Zhicai Wang and Y. Hao and Tingting Mu and Ouxiang Li and Shuo Wang and Xiangnan He},
  year   = {2026},
  eprint = {2303.08698},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/pdf/2303.08698},
  journal = {CVPR 2023 2023}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2303.08698.svg)](https://codeofpaper.com/paper/2303.08698)

Discussion