… In this section, we will first revisit the preliminary work on contrastive learning and address
its limitations. Then we will introduce our proposed weakly supervised contrastive learning …

… With this intuition, we present a two-stage weakly-supervised contrastive learning approach.
The first stage is to cluster data according to its auxiliary information. The second stage is to …

… Figure 1 provides an overview of our proposed contrastive training. We propose a novel
formulation of the weakly supervised phrase grounding problem as that of maximizing a lower …

… In this paper, we argue that learning by comparing helps identify these hard snippets and
we propose to utilize snippet Contrastive learning to Localize Actions, CoLA for short. …

… Weakly-supervised vision-language grounding aims to … Contrastive Learning (CCL) to
develop sufficient contrastive training … further explore weakly-supervised contrastive learning and …

… • We extend weakly-supervised contrastive learning to unsupervised object discovery for
the first time and introduce an alignment loss to keep the pixel-level semantic consistency for …

… In this article, we present weakly supervised contrastive learning (WSCL) and domain adaptation
(DA… Within each subdomain, contrastive learning using tracklet ids is applied to learn a …

… We introduce a weakly supervised contrastive learning method that allows us to consider
multiple positives and multiple negatives, and a prototype-based clustering method that avoids …

… Weakly supervised Contrastive Learning To regularize the training process, we propose a
weakly supervised contrastive loss (WCL). We first project the hidden representations of the …

… We formulate weakly supervised segmentation as a semi-supervised metric learning …
We propose 4 types of contrastive relationships between pixels and segments in the feature …