Abstract:We consider the critical issue of false negatives in Vision-Language Pre-training (VLP), a challenge that arises from the inherent many-to-many correspondence of image-text pairs in large-scale web-crawled datasets. The presence of false negatives can impede achieving optimal performance and even lead to learning failures. To address this challenge, we propose a method called COSMO (COnverting and SMOoothing false negatives) that manages the false negative issues, especially powerful in hard negative sampling. Building upon the recently developed GRouped mIni-baTch sampling (GRIT) strategy, our approach consists of two pivotal components: 1) an efficient connection mining process that identifies and converts false negatives into positives, and 2) label smoothing for the image-text contrastive loss (ITC). Our comprehensive experiments verify the effectiveness of COSMO across multiple downstream tasks, emphasizing the crucial role of addressing false negatives in VLP, potentially even surpassing the importance of addressing false positives. In addition, the compatibility of COSMO with the recent BLIP-family model is also demonstrated.
Abstract:Power control for the device-to-device interference channel with single-antenna transceivers has been widely analyzed with both model-based methods and learning-based approaches. Although the learning-based approaches, i.e., datadriven and model-driven, offer performance improvement, the widely adopted graph neural network suffers from learning the heterophilous power distribution of the interference channel. In this paper, we propose a deep learning architecture in the family of graph transformers to circumvent the issue. Experiment results show that the proposed methods achieve the state-of-theart performance across a wide range of untrained network configurations. Furthermore, we show there is a trade-off between model complexity and generality.
Abstract:Despite the recent development of deep learning-based point cloud upsampling, most MLP-based point cloud upsampling methods have limitations in that it is difficult to train the local and global structure of the point cloud at the same time. To solve this problem, we present a combined graph convolution and transformer for point cloud upsampling, denoted by PU-EdgeFormer. The proposed method constructs EdgeFormer unit that consists of graph convolution and multi-head self-attention modules. We employ graph convolution using EdgeConv, which learns the local geometry and global structure of point cloud better than existing point-to-feature method. Through in-depth experiments, we confirmed that the proposed method has better point cloud upsampling performance than the existing state-of-the-art method in both subjective and objective aspects. The code is available at https://github.com/dohoon2045/PU-EdgeFormer.