Think Bright, Diffuse Nice: Enhancing T2I-ICL via Inductive-Bias Hint Instruction and Query Contrastive Decoding

ACL ARR 2026 January Submission9747 Authors

Keywords: T2I-ICL, Training-free, Inference scaling, Prompt engineering, Decoding strategy

Abstract: Text-to-Image In-Context Learning (T2I-ICL) enables customized image synthesis via interleaved text-image examples but faces two mutually reinforcing bottlenecks, compliance failure and prior-dominated hallucination, that form a vicious cycle degrading generation quality. Existing methods rely on tailored training, which limits flexibility and raises deployment costs. To address these challenges effectively, we propose TBDN, a training-free framework integrating two complementary closed-loop mechanisms: Hint Instruction (HI) and Query Contrastive Decoding (QCD). HI injects task-aware inductive bias via lightweight prompt engineering to anchor models on contextual mapping rules, thereby mitigating compliance failure. QCD adjusts the decoding distributions of language models by contrasting full-input and query-omitted distributions, suppressing prior-dominated hallucination. TBDN achieves State-of-the-Art performance on CoBSAT and Text-to-Image Fast Mini-ImageNet, with robust generalization across model backbones, prompt designs, and hyperparameters. It also maintains promising performance in concept preservation and prompt following on Dreambench++. By breaking the two bottlenecks, TBDN establishes a simple yet effective framework for efficient and reliable T2I-ICL.

Paper Type: Long

Research Area: Multimodality and Language Grounding to Vision, Robotics and Beyond

Research Area Keywords: cross-modal content generation, cross-modal application

Languages Studied: english

Submission Number: 9747