Unverified50% confidenceFactExact time
The paper trains a bijection (typically implemented as a normalizing flow) so that the target distribution in the latent space approximates a reference density with simple structure.
1
Sources
50%
Confidence
Long-term
Relevance
10/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe viability of the asymmetric architecture depends on the two models being jointly trained or distilled so that their output vectors fall within the same metric space.66% similarVerified扩散模型以 DALL-E 2、Stable Diffusion、Midjourney 为代表，重新定义了图像生成任务66% similarUnverifiedCompared to score-function-based diffusion models, the flow matching training objective is more concise, and the trajectory can be designed as a rectified flow, enabling high-quality generation in fewer inference steps.65% similarUnverifiedTraditional Langevin dynamics relies on gradient information of the target distribution's density function and performs well when the density function is smooth and differentiable.65% similarUnverifiedA single-step generation in discrete diffusion can truly match the training data distribution only when the positions it writes to are mutually independent given the already-fixed tokens.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/990825API
curl https://kongchang.com/api/v1/knowledge/claims/990825MCP
get_claim(id=990825)