Research Thrusts | DeepThink Lab

Publications

For a complete list of publications, please visit Google Scholar.

Selected Publications

Compressible Dynamics in Deep Overparameterized Low-Rank Learning & Adaptation

Can Yaras, Peng Wang, Laura Balzano, Qing Qu.

International Conference on Machine Learning (ICML’24), 2024.

Oral Presentation (top 1.5%); Best Poster Award at MMLS’24

Preprint PDF BibTeX Code

The Emergence of Reproducibility and Generalizability in Diffusion Models

Huijie Zhang*, Jinfan Zhou*, Yifu Lu, Minzhe Guo, Liyue Shen, Qing Qu.

International Conference on Machine Learning (ICML’24), 2024.

Best Paper Award at NeurIPS’23 Workshop on Diffusion Models (news)

Preprint PDF BibTeX Slides Website

A Geometric Analysis of Neural Collapse with Unconstrained Features

Zhihui Zhu*, Tianyu Ding*, Jinxin Zhou, Xiao Li, Chong You, Jeremias Sulam, Qing Qu.

Neural Information Processing Systems (NeurIPS’21), 2021.

Spotlight Presentation (top 3%)

Preprint PDF Slides BibTeX Code Video

Analysis of the Optimization Landscapes for Overcomplete Representation Learning

Qing Qu, Yuexiang Zhai, Xiao Li, Yuqian Zhang, Zhihui Zhu.

International Conference on Learning Representations (ICLR’20), 2020.

Oral Presentation (top 1.9%)

Preprint PDF Slides BibTeX

Exact Recovery of Multichannel Sparse Blind Deconvolution via Gradient Descent

Qing Qu, Xiao Li, Zhihui Zhu.

SIAM Journal on Imaging Sciences, 13(3): 1630–1652, 2020. Preliminary version appeared at NeurIPS 2019.

Spotlight Presentation (top 3%)

Preprint PDF Code Poster Slides BibTeX

A Geometric Analysis of Phase Retrieval

Ju Sun, Qing Qu, John Wright.

Foundations of Computational Mathematics, 18(5): 1131-1198, 2018. Preliminary version appeared at ISIT 2016.

Preprint PDF Code Slides BibTeX

Preprint

Coarse-to-Fine Hierarchical Alignment for UAV-based Human Detection using Diffusion Models

Wenda Li, Meng Wu, Sungmin Eum, Heesung Kwon, Qing Qu.

Arxiv Preprint arXiv:2512.13869, 2025.

Preprint PDF

AlphaFlow: Understanding and Improving MeanFlow Models.

Huijie Zhang, Aliaksandr Siarohin, Willi Menapace, Michael Vasilkovsky, Sergey Tulyakov, Qing Qu, Ivan Skorokhodov.

Arxiv Preprint arXiv:2510.20771, 2025.

Preprint PDF Code

SpaceTools: Tool-Augmented Spatial Reasoning via Double Interactive RL.

Siyi Chen, Mikaela Angelina Uy, Chan Hee Song, Faisal Ladhak, Adithyavairavan Murali, Qing Qu, Stan Birchfield, Valts Blukis, Jonathan Tremblay.

Arxiv Preprint arXiv:2512.04069, 2025.

Preprint PDF

Unlearning Isn't Invisible: Detecting Unlearning Traces in LLMs from Model Outputs.

Yiwei Chen, Soumyadeep Pal, Yimeng Zhang, Qing Qu, Sijia Liu.

Arxiv Preprint arXiv:2506.14003, 2025.

Preprint PDF

Understanding Generalization in Diffusion Models via Probability Flow Distance

Huijie Zhang, Zijian Huang, Siyi Chen, Jinfan Zhou, Zekai Zhang, Peng Wang, Qing Qu.

Arxiv Preprint arXiv:2505.20123, 2025.

Preprint PDF

Out-of-Distribution Generalization of In-Context Learning: A Low-Dimensional Subspace Perspective

Soo Min Kwon*, Alec S. Xu*, Can Yaras, Laura Balzano, Qing Qu.

Arxiv Preprint arXiv:2505.14808, 2025.

Preprint PDF

The Dual Power of Interpretable Token Embeddings: Jailbreaking Attacks and Defenses for Diffusion Model Unlearning

Siyi Chen, Yimeng Zhang, Sijia Liu, Qing Qu.

Arxiv Preprint arXiv:2504.21307, 2025.

Preprint PDF

An Overview of Low-Rank Structures in the Training and Adaptation of Large Models

Laura Balzano, Tianjiao Ding, Benjamin D. Haeffele, Soo Min Kwon, Qing Qu, Peng Wang, Zhangyang Wang, Can Yaras.

Arxiv Preprint arXiv:2503.19859, 2025. (Authors are listed alphabetically)

Preprint PDF

Understanding How Nonlinear Layers Create Linearly Separable Features for Low-Dimensional Data

Alec S. Xu, Can Yaras, Peng Wang, Qing Qu.

Arxiv Preprint arXiv:2501.02364, 2025.

Preprint PDF BibTeX

Analyzing and Improving Model Collapse in Rectified Flow Models

Huminhao Zhu, Fangyikang Wang, Tianyu Ding, Qing Qu, Zhihui Zhu.

Arxiv Preprint arXiv:2412.08175, 2024.

Preprint PDF BibTeX

Diffusion Models Learn Low-Dimensional Distributions via Subspace Clustering

Peng Wang*, Huijie Zhang*, Zekai Zhang, Siyi Chen, Yi Ma, Qing Qu.

Arxiv Preprint arXiv:2409.02426, 2024.

Preprint PDF BibTeX Code Project Website

Decoupled Data Consistency with Diffusion Purification for Image Restoration

Xiang Li, Soo Min Kwon, Ismail R. Alkhouri, Saiprasad Ravishankar, Qing Qu.

ArXiv Preprint arXiv:2403.06054, 2024.

Preprint PDF BibTeX Code

2025

Towards Understanding the Mechanisms of Classifier-Free Guidance

Xiang Li, Rongrong Wang, Qing Qu.

Neural Information Processing Systems (NeurIPS'25), 2025.

Spotlight Presentation (top 3.2%)

Preprint PDF

A Closer Look at Model Collapse: From a Generalization-to-Memorization Perspective

Lianghe Shi*, Meng Wu*, Huijie Zhang, Zekai Zhang, Molei Tao, Qing Qu.

Neural Information Processing Systems (NeurIPS'25), 2025.

Spotlight Presentation (top 3.2%)

Preprint PDF Code Project Website

Shallow Diffuse: Robust and Invisible Watermarking through Low-Dimensional Subspaces in Diffusion Models

Wenda Li*, Huijie Zhang*, Qing Qu.

Neural Information Processing Systems (NeurIPS'25), 2025.

Spotlight Presentation (top 3.2%)

Preprint PDF BibTeX Code Project Website

Understanding Representation Dynamics of Diffusion Models via Low-Dimensional Modeling

Xiao Li*, Zekai Zhang*, Xiang Li, Siyi Chen, Zhihui Zhu, Peng Wang, Qing Qu.

Neural Information Processing Systems (NeurIPS'25), 2025.

Preprint PDF

FlowDAS: A Stochastic Interpolant-based Framework for Data Assimilation

Siyi Chen*, Yixuan Jia*, Qing Qu, He Sun, Jeffrey A. Fessler.

Neural Information Processing Systems (NeurIPS'25), 2025.

Preprint PDF BibTeX

UGoDIT: Unsupervised Group Deep Image Prior Via Transferable Weights

Shijun Liang, Ismail R. Alkhouri, Siddhant Gautam, Qing Qu, Saiprasad Ravishankar.

Neural Information Processing Systems (NeurIPS'25), 2025.

Preprint PDF BibTeX

Robust Physics-based Deep MRI Reconstruction Via Diffusion Purification

Ismail Alkhouri*, Shijun Liang*, Rongrong Wang, Qing Qu, Saiprasad Ravishankar.

IEEE Transactions on Neural Networks and Learning Systems (TNNLS), 2025.

Preprint PDF BibTeX

Understanding Deep Representation Learning via Layerwise Feature Compression and Discrimination

Peng Wang*, Xiao Li*, Can Yaras, Zhihui Zhu, Laura Balzano, Wei Hu, Qing Qu.

Journal of Machine Learning Research (JMLR), 2025.

Preprint PDF BibTeX Code

SITCOM: Step-wise Triple-Consistent Diffusion Sampling for Inverse Problems

Ismail Alkhouri, Shijun Liang, Cheng-Han Huang, Jimmy Dai, Qing Qu, Saiprasad Ravishankar, Rongrong Wang.

International Conference on Machine Learning (ICML'25), 2025.

Preprint PDF BibTeX

Attention-Only Transformers via Unrolled Subspace Denoising

Peng Wang, Yifu Lu, Yaodong Yu, Druv Pai, Qing Qu, Yi Ma.

International Conference on Machine Learning (ICML'25), 2025.

Preprint PDF

Explaining and Mitigating the Modality Gap in Contrastive Multimodal Learning

Can Yaras*, Siyi Chen*, Peng Wang, Qing Qu.

Conference on Parsimony and Learning (CPAL'25), 2025.

Preprint PDF BibTeX

Unfolding Videos Dynamics via Taylor Expansion

Siyi Chen, Minkyu Choi, Zesen Zhao, Kuan Han, Qing Qu, Zhongming Liu.

Conference on Parsimony and Learning (CPAL'25), 2025.

Preprint PDF

Learning Dynamics of Deep Matrix Factorization Beyond the Edge of Stability

Avrajit Ghosh*, Soo Min Kwon*, Rongrong Wang, Saiprasad Ravishankar, Qing Qu.

International Conference on Learning Representations (ICLR'25), 2025.

Preprint PDF BibTeX

Sequential Diffusion-Guided Deep Image Prior for Medical Image Reconstruction

Shijun Liang, Ismail Alkhouri, Qing Qu, Rongrong Wang, Saiprasad Ravishankar.

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP'25), 2025.

Preprint PDF BibTeX

Analysis of Deep Image Prior and Exploiting Self-Guidance for Image Reconstruction

Shijun Liang, Evan Bell, Qing Qu, Rongrong Wang, Saiprasad Ravishankar.

IEEE Transactions on Computational Imaging (TCI), 2025.

Preprint PDF BibTeX

2024

Understanding Generalizability of Diffusion Models Requires Rethinking the Hidden Gaussian Structure

Xiang Li, Yixiang Dai, Qing Qu.

Neural Information Processing Systems (NeurIPS'24), 2024.

Preprint PDF BibTeX

BLAST: Block-Level Adaptive Structured Matrices for Efficient Deep Neural Network Inference

Changwoo Lee, Soo Min Kwon, Qing Qu, Hun-Seok Kim.

Neural Information Processing Systems (NeurIPS'24), 2024.

Preprint PDF BibTeX Code

Image Reconstruction Via Autoencoding Sequential Deep Image Prior

Ismail Alkhouri, Shijun Liang, Evan Bell, Qing Qu, Rongrong Wang, Saiprasad Ravishankar.

Neural Information Processing Systems (NeurIPS'24), 2024.

Preprint BibTeX Code

Exploring Low-Dimensional Subspaces in Diffusion Models for Controllable Image Editing

Siyi Chen*, Huijie Zhang*, Minzhe Guo, Yifu Lu, Peng Wang, Qing Qu.

Neural Information Processing Systems (NeurIPS'24), 2024.

Preprint PDF BibTeX Code Project Website

Compressible Dynamics in Deep Overparameterized Low-Rank Learning & Adaptation

Can Yaras, Peng Wang, Laura Balzano, Qing Qu.

International Conference on Machine Learning (ICML'24), 2024.

Oral Presentation (top 1.5%)

Preprint PDF BibTeX Code

A Global Geometric Analysis of Maximal Coding Rate Reduction

Peng Wang, Huikang Liu, Druv Pai, Yaodong Yu, Zhihui Zhu, Qing Qu, Yi Ma.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX

Symmetric Matrix Completion with ReLU Sampling

Huikang Liu*, Peng Wang*, Longxiu Huang, Qing Qu, Laura Balzano.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX

Optimal Eye Surgeon: Finding Image Priors Through Sparse Generators at Initialization

Avrajit Ghosh, Xitong Zhang, Kenneth K. Sun, Qing Qu, Saiprasad Ravishankar, Rongrong Wang.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX Code

Sim2Real in Reconstructive Spectroscopy: Deep Learning with Augmented Device-Informed Data Simulation

Jiyi Chen*, Pengyu Li*, Yutong Wang, Pei-Cheng Ku, Qing Qu.

APL Machine Learning, Vol. 2, No. 3, pp. 036106, Aug. 2024.

Preprint PDF BibTeX

UV-VIS Chip-Scale Spectropolarimeter

Juhyeon Kim, Jiyi Chen, Pengyu Li, Yutong Wang, Qing Qu, Pei-Cheng Ku.

Proceedings of SPIE 13026, Next-Generation Spectroscopic Technologies XVI, 1302605, 2024.

Preprint PDF BibTeX

2023

Improving Efficiency of Diffusion Models via Multi-Stage Framework and Tailored Multi-Decoder Architectures

Huijie Zhang*, Yifu Lu*, Ismail Alkhouri, Saiprasad Ravishankar, Dogyoon Song, Qing Qu.

Conference on Computer Vision and Pattern Recognition (CVPR'24), 2024.

Preprint PDF BibTeX Code Project Website

Efficient Low-Dimensional Compression of Overparameterized Models

Soo Min Kwon*, Zekai Zhang*, Dogyoon Song, Laura Balzano, Qing Qu.

Proceedings of the 27th International Conference on Artificial Intelligence and Statistics (AISTATS'24), 2024.

Preprint PDF BibTeX Code

Neural Collapse in Multi-label Learning with Pick-all-label Loss

Pengyu Li*, Xiao Li*, Yutong Wang, Qing Qu.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX Code

Generalized Neural Collapse for a Large Number of Classes

Jiachen Jiang*, Jinxin Zhou*, Peng Wang, Qing Qu, Dustin Mixon, Chong You*, Zhihui Zhu*.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX

The Emergence of Reproducibility and Consistency in Diffusion Models

Huijie Zhang*, Jinfan Zhou*, Yifu Lu, Minzhe Guo, Liyue Shen, Qing Qu.

International Conference on Machine Learning (ICML'24), 2024.

Preprint PDF BibTeX Slides

Investigating the Catastrophic Forgetting in Multimodal Large Language Models

Yuexiang Zhai, Shengbang Tong, Xiao Li, Mu Cai, Qing Qu, Yong Jae Lee, Yi Ma.

Conference on Parsimony and Learning (CPAL'24), 2024.

Preprint PDF BibTeX

The Law of Parsimony in Gradient Descent for Learning Deep Linear Networks

Can Yaras*, Peng Wang*, Wei Hu, Zhihui Zhu, Laura Balzano, Qing Qu.

ArXiv Preprint arXiv:2306.01154, 2023.

Preprint PDF BibTeX Code Slides

Solving Inverse Problems with Latent Diffusion Models via Hard Data Consistency

Bowen Song*, Soo Min Kwon*, Zecheng Zhang, Xinyu Hu, Qing Qu, Liyue Shen.

International Conference on Learning Representations (ICLR'24), 2024.

Spotlight Presentation (top 5%)

Preprint PDF BibTeX

Robust Self-Guided Deep Image Prior

Evan Bell, Shijun Liang, Qing Qu, Saiprasad Ravishankar.

IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP'23), 2023.

Preprint PDF BibTeX

2022

Understanding and Improving Transfer Learning of Deep Models via Neural Collapse

Xiao Li*, Sheng Liu*, Jinxin Zhou, Xinyu Lu, Carlos Fernandez-Granda, Zhihui Zhu, Qing Qu.

Transactions on Machine Learning Research (TMLR), 2024.

Preprint PDF BibTeX

Neural Collapse with Normalized Features: A Geometric Analysis over the Riemannian Manifold

Can Yaras*, Peng Wang*, Zhihui Zhu, Laura Balzano, Qing Qu.

Neural Information Processing Systems (NeurIPS'22), 2022.

Preprint PDF BibTeX Code

Are All Losses Created Equal: A Neural Collapse Perspective

Jinxin Zhou, Chong You, Xiao Li, Kangning Liu, Sheng Liu, Qing Qu, Zhihui Zhu.

Neural Information Processing Systems (NeurIPS'22), 2022.

Preprint PDF BibTeX Code

Miniaturizing a Chip-Scale Spectrometer Using Local Strain Engineering and Total-Variation Regularized Reconstruction

Tuba Sarwar, Can Yaras, Xiang Li, Qing Qu, Pei-Cheng Ku.

Nano Letters, Vol. 22, No. 20, pp. 8174-8180, 2022.

Preprint PDF BibTeX

Hidden State Variability of Pretrained Language Models Can Guide Computation Reduction for Transfer Learning

Shuo Xie, Jiahao Qiu, Ankita Pasad, Li Du, Qing Qu, Hongyuan Mei.

Findings of Empirical Methods in Natural Language Processing (EMNLP), 2022.

Preprint PDF BibTeX Code

Robust Training under Label Noise by Over-parameterization

Sheng Liu, Zhihui Zhu, Qing Qu, Chong You.

International Conference on Machine Learning (ICML'22), 2022.

Preprint PDF BibTeX Code

On the Optimization Landscape of Neural Collapse under MSE Loss: Global Optimality with Unconstrained Features

Jinxin Zhou*, Xiao Li*, Tianyu Ding, Chong You, Qing Qu*, Zhihui Zhu*.

International Conference on Machine Learning (ICML'22), 2022.

Preprint PDF BibTeX Code

Linear Convergence Analysis of Neural Collapse with Unconstrained Features

Peng Wang*, Huikang Liu*, Can Yaras*, Laura Balzano, Qing Qu.

OPT 2022: Optimization for Machine Learning (NeurIPS 2022 Workshop), 2022.

Preprint PDF BibTeX

2021

Rank Overspecified Robust Matrix Recovery: Subgradient Method and Exact Recovery

Lijun Ding*, Liwei Jiang*, Yudong Chen, Qing Qu, Zhihui Zhu.

Neural Information Processing Systems (NeurIPS'21), 2021.

Preprint PDF BibTeX

A Geometric Analysis of Neural Collapse with Unconstrained Features

Zhihui Zhu*, Tianyu Ding*, Jinxin Zhou, Xiao Li, Chong You, Jeremias Sulam, Qing Qu.

Neural Information Processing Systems (NeurIPS'21), 2021.

Spotlight Presentation (top 3%)

Preprint PDF Slides BibTeX Code Video

Convolutional Normalization: Improving Deep Convolutional Network Robustness and Training

Sheng Liu*, Xiao Li*, Yuexiang Zhai, Chong You, Zhihui Zhu, Carlos Fernandez-Granda, Qing Qu.

Neural Information Processing Systems (NeurIPS'21), 2021.

Preprint PDF BibTeX Code

Weakly Convex Optimization over Stiefel Manifold Using Riemannian Subgradient-Type Methods

Xiao Li*, Shixiang Chen*, Zengde Deng, Qing Qu, Zhihui Zhu, Anthony Man Cho So.

SIAM Journal on Optimization, 31(3): 1605-1634, 2021.

Preprint PDF Code BibTeX

2020

Geometric Analysis of Nonconvex Optimization Landscapes for Overcomplete Learning

Qing Qu, Yuexiang Zhai, Xiao Li, Yuqian Zhang, Zhihui Zhu.

International Conference on Learning Representations (ICLR'20), 2020.

Oral Presentation (top 1.9%)

Preprint PDF Slides BibTeX

Robust Recovery via Implicit Bias of Discrepant Learning Rates for Double Over-parameterization

Chong You*, Zhihui Zhu*, Qing Qu, Yi Ma.

Neural Information Processing Systems (NeurIPS'20), 2020.

Spotlight Presentation (top 4%)

Preprint PDF BibTeX Code

Short and Sparse Deconvolution — A Geometric Approach

Yenson Lau*, Qing Qu*, Han-wen Kuo, Pengcheng Zhou, Yuqian Zhang, John Wright.

International Conference on Learning Representations (ICLR'20), 2020.

Preprint PDF Code Poster Slides Website BibTeX

From Symmetry to Geometry: Tractable Nonconvex Problems

Yuqian Zhang, Qing Qu, John Wright.

Preprint, 2020.

Preprint PDF Slides BibTeX

Finding the Sparsest Vectors in a Subspace: Theory, Algorithms, and Applications

Qing Qu*, Zhihui Zhu*, Xiao Li, Manolis C. Tsakiris, John Wright, Rene Vidal.

Preprint, 2020.

Preprint PDF Slides BibTeX

2019 & Prior

Exact Recovery of Multichannel Sparse Blind Deconvolution via Gradient Descent

Qing Qu, Xiao Li, Zhihui Zhu.

SIAM Journal on Imaging Sciences, 13(3): 1630-1652, 2020. Preliminary version appeared at NeurIPS 2019.

Spotlight Presentation (top 3%)

Preprint PDF Code Poster Slides BibTeX

Convolutional Phase Retrieval via Gradient Descent

Qing Qu, Yuqian Zhang, Yonina C. Eldar, John Wright.

IEEE Transactions on Information Theory, 66(3): 1785-1821, March 2020. Preliminary version appeared at NeurIPS 2017.

Preprint PDF Code Poster Slides BibTeX

A Geometric Analysis of Phase Retrieval

Ju Sun, Qing Qu, John Wright.

Foundations of Computational Mathematics, 18(5): 1131-1198, 2018. Preliminary version appeared at ISIT 2016.

Preprint PDF Code Slides BibTeX

Complete Dictionary Recovery over the Sphere I: Overview and the Geometric Picture

Ju Sun, Qing Qu, John Wright.

IEEE Transactions on Information Theory, 63(2): 853-884, February 2017. Preliminary version at ICML 2015.

Best Student Paper Award (SPARS 2015)

Preprint PDF Code Poster BibTeX

Complete Dictionary Recovery over the Sphere II: Recovery by Riemannian Trust-region Method

Ju Sun, Qing Qu, John Wright.

IEEE Transactions on Information Theory, 63(2): 885-914, February 2017.

Preprint PDF Code Poster BibTeX

Finding a Sparse Vector in a Subspace: Linear Sparsity Using Alternating Directions

Qing Qu, Ju Sun, John Wright.

IEEE Transactions on Information Theory, 62(10): 5855-5880, October 2016. Preliminary version appeared at NeurIPS 2014.

Preprint PDF Code Poster Slides BibTeX

Subspace Vertex Pursuit: A Fast and Robust Near-Separable Nonnegative Matrix Factorization Method for Hyperspectral Unmixing

Qing Qu, Nasser M. Nasrabadi, Trac D. Tran.

IEEE Journal of Selected Topics in Signal Processing, 9(6): 1142-1155, September 2015. Preliminary version appeared at ICASSP 2014.

PDF Poster BibTeX

When Are Nonconvex Problems Not Scary?

Ju Sun, Qing Qu, John Wright.

NeurIPS 2015 Workshop on Nonconvex Optimization for Machine Learning.

Preprint PDF BibTeX

Abundance Estimation for Bilinear Mixture Models via Joint Sparse and Low-Rank Representation

Qing Qu, Nasser M. Nasrabadi, Trac D. Tran.

IEEE Transactions on Geoscience and Remote Sensing, 52(7): 4404-4423, July 2014. Preliminary version appeared at ICASSP 2013.

PDF Code Poster Slides BibTeX

Structured Priors for Sparse-Representation-Based Hyperspectral Image Classification

Xiaoxia Sun, Qing Qu, Nasser M. Nasrabadi, Trac D. Tran.

IEEE Geoscience and Remote Sensing Letters, 11(7): 1235-1239, July 2014.

Preprint PDF BibTeX

Robust Zero-point Attraction Least Mean Square Algorithm on Near Sparse System Identification

Jian Jin, Qing Qu, Yuantao Gu.

IET Signal Processing, 7(3): 210-218, May 2013.

Preprint PDF BibTeX