HYPE: Hyperbolic Entailment Filtering for Underspecified Images and Texts

arXiv:2404.17507v1 Announce Type: new
Abstract: In an era where the volume of data drives the effectiveness of self-supervised learning, the specificity and clarity of data semantics play a crucial role in model training. Addressing this, we introduce HYPerbolic Entailment filtering (HYPE), a novel methodology designed to meticulously extract modality-wise meaningful and well-aligned data from extensive, noisy image-text pair datasets. Our approach leverages hyperbolic embeddings and the concept of entailment cones to evaluate and filter out samples with meaningless or underspecified …

abstract arxiv cs.cv data data semantics extract filtering hype images methodology novel role self-supervised learning semantics specificity supervised learning training type wise