dorsal/arxiv
View SchemaThink Bright, Diffuse Nice: Enhancing T2I-ICL via Inductive-Bias Hint Instruction and Query Contrastive Decoding
| Authors | Zhiyong Ma, Zhenpeng Li, Yuanjie Shi, Zhengping Li, Jiahao Chen, Qingyuan Chuai |
|---|---|
| Categories | |
| ArXiv ID | 2601.06169vv1 |
| URL | https://arxiv.org/abs/2601.06169 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
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.
{
"annotation_id": "330d3562-afbd-4082-accf-99e6ec2900e8",
"date_created": "2026-02-17T05:53:08.753000Z",
"date_modified": "2026-02-17T05:53:08.753000Z",
"file_hash": "24789263aee41edcba0ed40bd03f98b77fa5e88b2bb4c83575c783297cf48980",
"private": false,
"record": {
"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.",
"arxiv_id": "2601.06169",
"authors": [
"Zhiyong Ma",
"Zhenpeng Li",
"Yuanjie Shi",
"Zhengping Li",
"Jiahao Chen",
"Qingyuan Chuai"
],
"categories": [
"cs.CV"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "Think Bright, Diffuse Nice: Enhancing T2I-ICL via Inductive-Bias Hint Instruction and Query Contrastive Decoding",
"url": "https://arxiv.org/abs/2601.06169",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "1f060e3d-164b-47c7-a0ce-65e92c0b7fc1",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}