no code implementations • 7 Apr 2024 • Yimu Wang, Shuai Yuan, Xiangru Jian, Wei Pang, Mushi Wang, Ning Yu
While recent progress in video-text retrieval has been driven by the exploration of powerful model architectures and training strategies, the representation learning ability of video-text retrieval models is still limited due to low-quality and scarce training data annotations.
1 code implementation • 20 Oct 2023 • Xiangru Jian, Yimu Wang
However, a recent study shows that multi-modal data representations tend to cluster within a limited convex cone (as representation degeneration problem), which hinders retrieval performance due to the inseparability of these representations.
1 code implementation • 17 Oct 2023 • Yimu Wang, Xiangru Jian, Bo Xue
In this work, we present a post-processing solution to address the hubness problem in cross-modal retrieval, a phenomenon where a small number of gallery data points are frequently retrieved, resulting in a decline in retrieval performance.
no code implementations • 18 Aug 2022 • Qixin Zhang, Zengde Deng, Xiangru Jian, Zaiyi Chen, Haoyuan Hu, Yu Yang
Maximizing a monotone submodular function is a fundamental task in machine learning, economics, and statistics.
no code implementations • 11 Oct 2018 • Xiangru Jian, Paulo J. M Monteiro, Kimberly E. Kurtis
This paper mainly describes the development of a new type of regression model to predict the long-term expansion of concrete subjected to a sulfate-rich environment.
Applications