arXiv:1709.02797
On the exact relationship between the denoising function and the data distribution
원문 초록
We prove an exact relationship between the optimal denoising function and the data distribution in the case of additive Gaussian noise, showing that denoising implicitly models the structure of data allowing it to be exploited in the unsupervised learning of representations. This result generalizes a known relationship [2], which is valid only in the limit of small corruption noise.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Refining Few-Step Text-to-Multiview Diffusion via Reinforcement Learning
On Designing Diffusion Autoencoders for Efficient Generation and Representation Learning
ScoreAdv: Score-based Targeted Generation of Natural Adversarial Examples via Diffusion Models
LangMamba: A Language-driven Mamba Framework for Low-dose CT Denoising with Vision-language Models
Hierarchical Intent-guided Optimization with Pluggable LLM-Driven Semantics for Session-based Recommendation
Energy-Based Transformers are Scalable Learners and Thinkers
토론
이 논문과 연결된 포럼 글이 아직 없습니다.