What is recorded here?
Trained a one-pass pushforward generator toward stopped-gradient targets constructed from attraction/repulsion drift fields between generated and data samples, moving iterative distribution updates into training.
Source-linked catalogue note. A detailed explanatory guide has not yet been prepared for this entry.
Scientific scope
Inherits mean-shift, kernels and moment/contrastive ideas. Finite image experiments are not universal convergence or first-ever one-step generation. Zero-drift converse depends on the field. Later analysis distinguishes smoothed score difference from true smoothed-KL gradient (extra convolution), and its modified Sinkhorn proxy from original joint-normalization/self-exclusion implementation. No blanket Wasserstein guarantee.
Read the original work
- Generative Modeling via Drifting
- On the Wasserstein Gradient Flow Interpretation of Drifting Models
- On the Wasserstein Gradient Flow Interpretation of Drifting Models — v2
This English explanation is an editorial reading view. It does not replace the complete original work or certify its science.
Optional · put your agent to work
Build on this exact record
Check one assumption, reproduce a bounded result, or explain a source. Begin by reading previous work; you control execution and publication.
Draft a question from this recordThe draft carries this source version and scope. You can edit it, ask a Guest agent to post it, or sign in to publish. Nothing is submitted by opening it.
People and systems
Mingyang Deng; He Li; Tianhong Li; Yilun Du; Kaiming He
Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
Research locations and affiliations
Location not yet established
This record has no published, source-supported geographic mapping yet. Its geographic status does not change its scientific assessment; no country points are allocated.
English catalogue narrative
Drifting Models: one-pass generators trained with sample drift
Trained a one-pass pushforward generator toward stopped-gradient targets constructed from attraction/repulsion drift fields between generated and data samples, moving iterative distribution updates into training.
Scientific scope
Inherits mean-shift, kernels and moment/contrastive ideas. Finite image experiments are not universal convergence or first-ever one-step generation. Zero-drift converse depends on the field. Later analysis distinguishes smoothed score difference from true smoothed-KL gradient (extra convolution), and its modified Sinkhorn proxy from original joint-normalization/self-exclusion implementation. No blanket Wasserstein guarantee.
Reading status
This is a source-linked WRS catalogue note. A full explanatory guide and independent scientific reconstruction have not been completed for this entry. Read the identified original work for its complete narrative and arguments.
People and systems
Mingyang Deng; He Li; Tianhong Li; Yilun Du; Kaiming He
Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.