arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2103.16173cs.CVcs.AI
收藏

广义零样本学习的对比嵌入

Contrastive Embedding for Generalized Zero-Shot Learning

Nanjing University of Science and Technology(南京理工大学);RIKEN Center for Advanced Intelligence Project(日本理化学研究所先进智能研究中心)

Zongyan Han, Zhenyong Fu, Shuo Chen, Jian Yang

摘要:

广义零样本学习(GSRL)的目的是识别对象从可见和不可见的类,当只有标记的例子从可见的类提供。最近的特征生成方法学习了一个生成模型,可以合成看不见的类的缺失视觉特征,以减轻GALML中的数据不平衡问题。然而,原始的视觉特征空间是次优的GWML分类,因为它缺乏判别信息。为了解决这个问题,我们建议将生成模型与嵌入模型相结合,从而产生一个混合的GREML框架。混合GSTML方法将生成模型产生的真实和合成样本映射到嵌入空间中,在那里我们执行最终的GSTML分类。具体来说,我们提出了一个对比嵌入(CE)为我们的混合GREML框架。所提出的对比嵌入不仅可以利用类的监督,也可以利用实例的监督,后者通常被忽略了现有的GALML研究。我们评估了我们提出的混合GALML框架与对比嵌入,命名为CE-GALML,在五个基准数据集。结果表明,我们的CEGSTNL方法可以在三个数据集上以显着的幅度优于最先进的方法。我们的代码可在https://github.com/Hanzy1996/CE-GZSL上获得。

英文摘要:

Generalized zero-shot learning (GZSL) aims to recognize objects from both seen and unseen classes, when only the labeled examples from seen classes are provided. Recent feature generation methods learn a generative model that can synthesize the missing visual features of unseen classes to mitigate the data-imbalance problem in GZSL. However, the original visual feature space is suboptimal for GZSL classification since it lacks discriminative information. To tackle this issue, we propose to integrate the generation model with the embedding model, yielding a hybrid GZSL framework. The hybrid GZSL approach maps both the real and the synthetic samples produced by the generation model into an embedding space, where we perform the final GZSL classification. Specifically, we propose a contrastive embedding (CE) for our hybrid GZSL framework. The proposed contrastive embedding can leverage not only the class-wise supervision but also the instance-wise supervision, where the latter is usually neglected by existing GZSL researches. We evaluate our proposed hybrid GZSL framework with contrastive embedding, named CE-GZSL, on five benchmark datasets. The results show that our CEGZSL method can outperform the state-of-the-arts by a significant margin on three datasets. Our codes are available on https://github.com/Hanzy1996/CE-GZSL.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑