Unverified50% confidenceFactExact time
扩散模型于2020年前后由斯坦福大学团队正式引入图像生成领域，并在随后数年内取代GAN成为主流文生图范式
1
Sources
50%
Confidence
Long-term
Relevance
7/10/2026
First Seen
Sources
提示词工程入门：原理、方法与实战案例全解析
bilibili大模型零基础入门_6/30/2026
Related Claims
UnverifiedDiffusion models rose to prominence around 2020 and replaced GANs as the mainstream paradigm for image generation, thanks to more stable training processes and greater generation diversity.76% similarVerifiedAfter 2022, diffusion models comprehensively surpassed GANs (Generative Adversarial Networks), which had dominated for many years, in image generation quality, becoming the mainstream paradigm of generative AI.75% similarUnverifiedDiffusion models emerged around 2020, represented by OpenAI's GLIDE, Stability AI's Stable Diffusion, and Google's Imagen, and have gradually replaced GANs as the dominant paradigm in the field of image generation.75% similarVerifiedThe term "foundation model" was formally introduced by Stanford University's HAI Institute in 2021.74% similarVerifiedStanford University's DDPM paper in 2020 introduced diffusion models into the field of image generation.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/465106API
curl https://kongchang.com/api/v1/knowledge/claims/465106MCP
get_claim(id=465106)