Yexin Wang
10 Papers
1 Citations
Yexin Wang is an academic researcher. The author has contributed to research in topics: Computer science & Chemistry. The author has an hindex of 5, co-authored 10 publications.
Chat about Author
Papers
Metadata-Induced Contrastive Learning for Zero-Shot Multi-Label Text Classification
Yu Yvette Zhang,Zhihong Shen,Chieh-Han Wu,Boya Xie,Junheng Hao,Yexin Wang,Kuansan Wang,Jiawei Han +7 more
- 11 Feb 2022
TL;DR: Experimental results show that MICoL significantly outperforms strong zero-shot text classification and contrastive learning baselines and is on par with the state-of-the-art supervised metadata-aware LMTC method trained on 10K–200K labeled documents, and tends to predict more infrequent labels than supervised methods, thus alleviates the deteriorated performance on long-tailed labels.
Towards Universal Backward-Compatible Representation Learning
Binjie Zhang,Yixiao Ge,Yantao Shen,Shupeng Su,Chun Yuan,Xuyuan Xu,Yexin Wang,Ying Shan +7 more
- 03 Mar 2022
TL;DR: This work introduces a new problem of universal backward-compatible representation learning, covering all possible data split in model upgrades, and proposes a simple yet effective method, dubbed as Universal Backward-Compatible Training (UniBCT), to learn compatible representations in all kinds of model upgrading benchmarks in a unified manner.
13
Journal Article
Hot-Refresh Model Upgrades with Regression-Alleviating Compatible Training in Image Retrieval
TL;DR: A Regression-Alleviating Compatible Training (RACT) method to properly constrain the feature compatibility while reducing negative flips and an efficient uncertainty-based backfilling strategy is further introduced to fasten accuracy improvements.
7
Knowledge-augmented Few-shot Visual Relation Detection
Tianyuan Yu,Yang Liu,Jiaoyan Chen,Yinghui Li,Haitao Zheng,Xi Chen,Qingbin Liu,Wenqiang Liu,Dongxia Huang,Bei Wu,Yexin Wang +10 more
TL;DR: Zhang et al. as discussed by the authors proposed a knowledge-augmented, few-shot VRD framework leveraging both textual knowledge and visual relation knowledge to improve the generalization ability of few shot VRD.
PTVD: A Large-Scale Plot-Oriented Multimodal Dataset Based on Television Dramas
TL;DR: The plot-oriented multimodal dataset (PTVD) as discussed by the authors is the first non-English dataset of its kind, containing 1,106 TV drama episodes and 24,875 informative plot-focused sentences written by professionals, with the help of 449 human annotators.
2