SatisDive improves worst-candidate reward over FK steering by up to 0.43 with FLUX.1-dev and HPSv3
The paper introduces SatisDive, a training-free inference-time method for text-to-image generation that balances reward and diversity by treating generation as a satisficing problem. It ensures each image meets a reward floor while the batch satisfies a diversity cutoff. On the Pick-a-Pic dataset, SatisDive improves worst-candidate reward over FK steering by up to 0.43 when using FLUX.1-dev as…
1 source primary source
Key points
- SatisDive is a training-free inference-time method for text-to-image generation
- It improves worst-candidate reward by up to 0.43 with FLUX.1-dev and HPSv3 on Pick-a-Pic
- It improves worst-candidate reward by up to 0.70 with SANA-1.6B and ImageReward on Pick-a-Pic
Read the original at arXiv cs.AI · by Kevin Zhai, Siva Rajesh Kasa, Soumya Roy, Sumit Negi, Mubarak Shah primary sourceOpen source ↗
Topics · follow one to build your own front page
FLUX.1-devSANA-1.6B
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Google VP Yossi Matias says AI’s biggest impact may come from intersecting fields · 1 src
- Researchers adapt speech language model for simultaneous translation using prefix supervision · 1 src
- Study finds inductive prompting most consistent for LLM generalization in temporal extraction · 1 src
- AraBERT-based framework reaches 96.88% accuracy on Arabic DP ambiguity · 1 src
- Researchers introduce APDMem hierarchical memory for long-context LLM assistants · 1 src
Comments
via GitHub Discussions