Aditya ChinchurePaper Summary — A Generalization of Transformer Networks to GraphsThe Graph Transformer is a method that modifies the attention mechanism of the original transformer to better train on graphs.Jun 5, 2021Jun 5, 2021
Aditya ChinchurePaper Summary — BASNet: Boundary-Aware Salient Object DetectionBASNet is a method for salient object detection and segmentation based on the simple yet effective UNet architecture, with a focus on…Jun 1, 2021Jun 1, 2021
Aditya ChinchurePaper Summary — Referring Image Segmentation via Cross-Modal Progressive ComprehensionCMPC incorporates a fully-connected spatial graph and GCN to localize the object in the image, and then uses convLSTM to merge features…May 27, 2021May 27, 2021
Aditya ChinchurePaper Summary — ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for…ViLBERT (Lu et al. 2019) stands for Vision-and-Language BERT. ViLBERT is a multi-task model for multimodal tasks, based on the Transformer.May 26, 2021May 26, 2021