跳到正文
HuggingFace Daily Papers(社區熱門論文)·· 2026-10-15AI 評分37

SCOPE:面向文字生成圖像提示詞工程的語義邊界信用最佳化

SCOPE: Semantic Bounded Credit Optimization for Prompt Engineering in Text-to-image Generation

AI 導讀

SCOPE提出基於GRPO的強化學習目標,按語義片段調整策略更新權重,以改寫文字生成圖像提示詞並提升生成品質,無需額外圖像渲染。在 Qwen-Image、SD3 和 FLUX.1-dev 上,相較 vanilla GRPO,SCOPE 在 GeneVal 和 T2I-CompBench 的平均相對提升分別為 9.48% 和 7.62%。

正文

Published on Oct 15

·

Submitted by

Huang

on Oct 16

Authors:

,

,

,

,

,

,

Abstract

SCOPE improves text-to-image prompt rewriting via reinforcement learning by aligning policy updates with semantic spans, boosting generation quality without extra rendering.

Reinforcement learning (RL) enables language models to improve text-to-image (T2I) generation by rewriting prompts for a frozen image generator. Existing methods typically apply a single signal derived from the image to individual subword tokens, although objects, counts, attributes, and relations form coherent semantic spans. We propose SCOPE, an objective based on Group Relative Policy Optimization (GRPO) that aligns update weighting with this structure. The rewriter jointly generates an enhanced prompt and its semantic partition. SCOPE pools shifts in policy log probabilities within each span to derive a shared coefficient for its token losses while retaining a single advantage for each completion. It requires neither separate span rewards nor additional image renders. Across Qwen-Image, SD3, and FLUX.1-dev, SCOPE achieves average relative gains over vanilla GRPO of 9.48% on GeneVal and 7.62% on T2I-CompBench. Ablation studies show that both semantic boundaries and pooling within spans contribute to these gains, supporting semantic structure as a practical inductive bias for prompt engineering with RL.

View arXiv page View PDF GitHub Add to collection

Get this paper in your agent:

hf papers read 2610.22264

Don't have the latest CLI?

curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2610.22264 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2610.22264 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2610.22264 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.

來源:HuggingFace Daily Papers(社區熱門論文) · huggingface.co