Prompt-Driven Feature Diffusion for Open-World Semi-Supervised Learning

arXiv:2404.11795v1 Announce Type: new
Abstract: In this paper, we present a novel approach termed Prompt-Driven Feature Diffusion (PDFD) within a semi-supervised learning framework for Open World Semi-Supervised Learning (OW-SSL). At its core, PDFD deploys an efficient feature-level diffusion model with the guidance of class-specific prompts to support discriminative feature representation learning and feature generation, tackling the challenge of the non-availability of labeled data for unseen classes in OW-SSL. In particular, PDFD utilizes class prototypes as prompts in the diffusion model, …

abstract arxiv class core cs.ai cs.cv cs.lg diffusion diffusion model feature framework guidance novel open-world paper prompt prompts representation representation learning semi-supervised semi-supervised learning ssl supervised learning support type world