My name is Qian Qiao (乔谦). I am currently a researcher at Soul AILab. I received M.Sc. and B.Eng. from Soochow University, advised by Professors Fanzhang Li. My previous research focused on multimodal understanding, image generation, text spotting, and few-shot learning. Additionally, I have four years of experience in blockchain technology and investment.

Currently, I also lead a research team, focus on: diffusion language models, real-time video generation, and multimodal models. And I am actively seeking collaborators in these fields!

🔥 News

• 2025.01: 🎉 QPruner:ProbabilisticDecision Quantization for StructuredPruning in Large Language Models is accepted by NAACL 2025.
• 2024.12: 🎉 AIM: Let Any Multimodal Large Language Models Embrace Efficient In-Context Learning is accepted by AAAI 2024.
• 2024.11: 🎉 Invited by CogSci, ICME and IJCNN as Reviewer.
• 2024.10: 🎉 🔥🔥🔥 I honoured the national scholarship.
• 2024.07: 🎉 Invited by AAAI as Reviewer.
• 2024.08: 🎉 One paper is accepted by PRICAI 2024.
• 2024.07: 🎉 🔥🔥🔥 DNTextSpotter: Arbitrary-Shaped Scene Text Spotting via Improved Denoising Training is accepted by ACM MM 2024.
• 2024.07: 🎉 Invited by EMNLP as Reviewer.
• 2024.06: 🎉 Two papers are accepted by ICANN 2024.
• 2024.06: 🎉 Two papers are submitted to EMNLP 2024 as co-author.
• 2024.05: 🎉 Invited by NeurIPS as Reviewer.
• 2024.04: 🎉 Attending ICASSP2024 in Seoul, South Korea.
• 2024.04: 🎉 Two papers are accepted by ICME 2024.
• 2024.03: 🎉 Invited by MM as Program Reviewer.

📝 Publications

EuroGraphics 2026

• TextFlux: An OCR-Free DiT Model for High-Fidelity Multilingual Scene Text Synthesis,

  EuroGraphics, 2026.

  [arxiv] [code] [project]

Technical Report

• Marrying Autoregressive Transformer and Diffusion with Multi-Reference Autoregression

  Technical Report, 2025.

  [arxiv] [code]

ACM MM 2024

• DNTextSpotter: Arbitrary-Shaped Scene Text Spotting via Improved Denoising Training,

  ACM MM, 2024.

  [arxiv] [code] [project]

AAAI 2024

• AIM: Let Any Multi-modal Large Language Models Embrace Efficient In-Context Learning

  AAAI, 2025.

  

ICASSP 2024

• Talds-net: Task-aware adaptive local descriptors selection for few-shot image classification

  ICASSP, 2024.

  [arxiv] [code]

Neurocomputing 2025

• A Lie group Laplacian Support Vector Machine for semi-supervised learning

  Neurocomputing, 2025.

  [arxiv] [code]

PRICAI 2025

• Deeptts: Enhanced transformer-based text spotter via deep interaction between detection and recognition tasks

  PRICAI, 2024.

  [arxiv] [code]

NAACL 2025

• QPruner: Probabilistic Decision Quantization for Structured Pruning in Large Language Models

  NAACL, 2024.

  [arxiv] [code]

Other Publications

• Large Language Model Compression with Global Rank and Sparsity Optimization
• BVRCC: Bootstrapping Video Retrieval via Cross-Matching Correction
• Temporal-Conditional Referring Video Object Segmentation with Noise-Free Text-to-Video Diffusion Model
• STRA: A Simple Token Replacement Strategy Alleviating Exposure Bias in Text Generation
• Efficient fine-tuning of quantized models via adaptive rank and bitwidth

🎖 Honors and Awards

• National Scholarship, 2024 (MS. student)
• Outstanding Graduate Award, 2025 (Rate<5%)
• Special Grade Scholarship, 2023 (MS. student)

🆚 Academic Competitions

• 2024.06: My collaborator Yu Xie and I won three first places and one second place in the ICADR2024-Text Map Challenge (ICADR is one of the most authoritative conferences in the field of OCR), and we have been invited to present a technical report at ICADR2024.

💻 Internships

• 2024.02 - 2024.05, Research Intern, bilibili Inc, Shanghai, China.
• 2024.06 - 2024.10, Research Intern, Soul AILab, Shanghai, China.

Visitors