Publications

You can also find my articles on my Google Scholar profile.

Pre-Prints

Make Optimization Once and for All with Fine-grained Guidance

Published in Arxiv, 2025

Recommended citation: Make Optimization Once and for All with Fine-grained Guidance. Arxiv. M. Shi, R. Lin, X. Chen, Y. Zhou, Z. Ding, P. Li, T. Wang, K. Wang, Z. Wang, J. Zhang, T. Chen.
Download Paper

Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training

Published in Arxiv, 2024

Recommended citation: Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training. Arxiv. M. Shi*, Y. Zhou*, R. Yu, Z. Li, Z. Liang, X. Zhao, X. Peng, T Rajpurohit, R. Vedantam, W. Zhao, K. Wang, Y. You.
Download Paper

Conference Papers

Saliency-Aware Multi-Route Thinking: Grounding and Reasoning on Vision-Language Agents

Published in NeurIPS, 2026

Recommended citation: M. Shi, Y. He, Y. Zhu, J. Li. Saliency-Aware Multi-Route Thinking: Grounding and Reasoning on Vision-Language Agents. NeurIPS 2026.
Download Paper

Dive into the Implicit Biases of Low-rank Vision-language Alignment

Published in ECCV 2026, 2026

Recommended citation: Dive into the Implicit Biases of Low-rank Vision-language Alignment. ECCV 2026. M. Shi, S. Wang, X. Wang, S. Zhou, K. Wang, T. Fu, C. Zhao, A. Su, P. Jiang, M. Wu
Download Paper

A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training

Published in CVPR, 2025

Recommended citation: A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training. CVPR 2025. K. Wang*, M. Shi*, Y. Zhou*,Z. Yuan, Y. Shang, X. Peng, H. Zhang, Y. You
Download Paper

PRIOR: Personalized Prior for Reactivating the Information Overlooked in Federated Learning

Published in NeurIPS 2023, 2023

Recommended citation: PRIOR: Personalized Prior for Reactivating the Information Overlooked in Federated Learning. NeurIPS 2023. M. Shi, Y. Zhou, K. Wang, H. Zhang, S. Huang, Q. Ye, J. Lv
Download Paper

Unconstrained Feature Model and Its General Geometric Patterns in Federated Learning: Local Subspace Minority Collapse

Published in ICONIP 2023, 2023

Recommended citation: Unconstrained Feature Model and Its General Geometric Patterns in Federated Learning: Local Subspace Minority Collapse. ICONIP 2023. M. Shi, Y. Zhou, Q. Ye, J. Lv
Download Paper

DD-Ranking: Rethinking the Evaluation of Dataset Distillation

Published in NeurIPS (Oral), 2026

Recommended citation: DD-Ranking: Rethinking the Evaluation of Dataset Distillation. NeurIPS 2026 (Oral). Z. Li, X. Zhong, S. Khaki, Z. Liang, Y. Zhou, M. Shi, Z. Wang, X. Zhao, W. Zhao, Z. Qin, M. Wu, P. Zhou, H. Wang, D. J. Zhang, J. Liu, S. Wang, D. Liu, L. Zhang, G. Li, K. Wang, Z. Zhu, Z. Ma, J. T. Zhou, J. Lv, Y. Jin, P. Wang, K. Zhang, L. Lyu, Y. Huang, Z. Akata, Z. Deng, X. Wu, G. Cazenavette, Y. Shang, J. Cui, J. Gu, Q. Zheng, H. Ye, S. Wang, X. Wang, Y. Yan, A. Yao, M. Z. Shou, T. Chen, H. Bilen, B. Mirzasoleiman, M. Kellis, K. N. Plataniotis, Z. Wang, B. Zhao, Y. You, K. Wang
Download Paper

Grounding Memory Summarization in Utility Intent

Published in EMNLP, 2026

Recommended citation: Z. Lei, M. Shi, X. Fu, H. He, Q. R. Wang, J. Li. Grounding Memory Summarization in Utility Intent. EMNLP 2026.
Download Paper

IAPO: Information-Aware Policy Optimization for Token-Efficient Reasoning

Published in ICML 2026, 2026

Recommended citation: IAPO: Information-Aware Policy Optimization for Token-Efficient Reasoning. ICML 2026. Y. He, Y. Zhu, M. Shi, W. Zheng, L. Su, X. Wang, Q. Guo, J. Li
Download Paper

FedSH: Tackling Staleness By Scheduling High-order Approximation in Asynchronous Federated Learning

Published in IJCNN 2025, 2025

Recommended citation: FedSH: Tackling Staleness By Scheduling High-order Approximation in Asynchronous Federated Learning. IJCNN 2025. H. Gao, M. Shi, Y. Zhou, D. Xiong, Q. Ye, J. Lv
Download Paper

Pruning-Robust Mamba with Asymmetric Multi-Scale Scanning Paths

Published in NeurIPS 2025, 2025

Recommended citation: Pruning-Robust Mamba with Asymmetric Multi-Scale Scanning Paths. NeurIPS 2025. J. Lv, Y. Zhou, M. Shi, Z. Liang, P. Zhang, X. Peng, W. Zhao, Z. Zhu, J. Lv, Q. Ye, K. Wang
Download Paper

Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights

Published in NeurIPS 2025, 2025

Recommended citation: Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights. NeurIPS 2025. Z. Liang, D. Tang, Y. Zhou, X. Zhao, M. Shi, W. Zhao, Z. Li, P. Wang, K. Schürholt, D. Borth, M. M. Bronstein, Y. You, Z. Wang, K. Wang
Download Paper

Ferret: An Efficient Online Continual Learning Framework under Varying Memory Constraints

Published in CVPR, 2025

Recommended citation: Ferret: An Efficient Online Continual Learning Framework under Varying Memory Constraints. CVPR 2025. Y. Zhou, Y. Tian, J. Lv, M. Shi, Y. Li, Q. Ye, S. Zhang, J. Lv
Download Paper

REPA Works Until It Doesn’t: Early-Stopped, Holistic Alignment Supercharges Diffusion Training

Published in NeurIPS 2025, 2025

Recommended citation: REPA Works Until It Does not: Early-Stopped, Holistic Alignment Supercharges Diffusion Training. NeurIPS 2025. Z. Wang, W. Zhao, Y. Zhou, Z. Li, Z. Liang, M. Shi, X. Zhao, P. Zhou, K. Zhang, Z. Wang, K. Wang, Y. You
Download Paper

GSQ-Tuning: Group-Shared Exponents Integer in Fully Quantized Training for LLMs On-Device Fine-tuning

Published in ACL Findings, 2025

Recommended citation: GSQ-Tuning: Group-Shared Exponents Integer in Fully Quantized Training for LLMs On-Device Fine-tuning. Arxiv. S. Zhou*, S. Wang*, Z. Yuan*, M. Shi, Y. Shang, D. Yang.
Download Paper

Tackling Feature-Classifier Mismatch in Federated Learning via Prompt-Driven Feature Transformation

Published in NeurIPS 2025, 2024

Recommended citation: Tackling Feature-Classifier Mismatch in Federated Learning via Prompt-Driven Feature Transformation. NeurIPS 2025. X. Wu, J. Niu, X. Liu, M. Shi, G. Zhu, S. Tang.
Download Paper

Federated CINN Clustering for Accurate Clustered Federated Learning

Published in ICASSP 2024, 2024

Recommended citation: Federated CINN Clustering for Accurate Clustered Federated Learning. ICASSP 2024. Y. Zhou, M. Shi, Y. Tian, Y. Li, Q. Ye, J. Lv
Download Paper

Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence.

Published in ICCV 2023, 2023

Recommended citation: Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence. ICCV 2023. Y. Zhou, M. Shi, Y. Li, Y. Sun, Q. Ye, J. Lv
Download Paper

Mingjia Shi

Publications

Pre-Prints

Make Optimization Once and for All with Fine-grained Guidance

Faster Vision Mamba is Rebuilt in Minutes via Merged Token Re-training

Conference Papers

Saliency-Aware Multi-Route Thinking: Grounding and Reasoning on Vision-Language Agents

Dive into the Implicit Biases of Low-rank Vision-language Alignment

A Closer Look at Time Steps is Worthy of Triple Speed-Up for Diffusion Model Training

PRIOR: Personalized Prior for Reactivating the Information Overlooked in Federated Learning

Unconstrained Feature Model and Its General Geometric Patterns in Federated Learning: Local Subspace Minority Collapse

DD-Ranking: Rethinking the Evaluation of Dataset Distillation

Grounding Memory Summarization in Utility Intent

IAPO: Information-Aware Policy Optimization for Token-Efficient Reasoning

FedSH: Tackling Staleness By Scheduling High-order Approximation in Asynchronous Federated Learning

Pruning-Robust Mamba with Asymmetric Multi-Scale Scanning Paths

Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights

Ferret: An Efficient Online Continual Learning Framework under Varying Memory Constraints

REPA Works Until It Doesn’t: Early-Stopped, Holistic Alignment Supercharges Diffusion Training

GSQ-Tuning: Group-Shared Exponents Integer in Fully Quantized Training for LLMs On-Device Fine-tuning

Tackling Feature-Classifier Mismatch in Federated Learning via Prompt-Driven Feature Transformation

Federated CINN Clustering for Accurate Clustered Federated Learning

Communication-efficient Federated Learning with Single-Step Synthetic Features Compressor for Faster Convergence.

Journal Articles

E-3SFC: Communication-Efficient Federated Learning With Double-Way Features Synthesizing

DeFTA: A Plug-and-Play Peer-to-Peer Decentralized Federated Learning Framework

DLB: a dynamic load balance strategy for distributed training of deep neural networks

FLSGD: free local SGD with parallel synchronization