Recovery Guarantees of the Matrix Lasso Problem: On Equivalence of Convex and Non-Convex Formulations
Preprint.
Jinji Yang, Xinyuan Song, Chungen Shen, Ziye Ma
Download here
Publications
VAMO: Efficient Large-Scale Nonconvex Optimization via Adaptive Zeroth Order Variance Reduction
Preprint.
Jiahe Chen, Ziye Ma
Download here
Towards Optimal Branching of Linear and Semidefinite Relaxations for Neural Network Robustness Certification
Journal of Machine Learning Research (JMLR)
Brendon G Anderson, Ziye Ma, Jingqi Li, Somayeh Sojoudi
Download here
Deterministic Escape from Local Minima: Oracle from Simulated Over-parametrization
Preprint.
Tianqi Shen, Kunhan Gao, Ziye Ma
Download here
Structured Noise to Help Non-Convexity: Solving Matrix Completion as Noisy Matrix Sensing
Preprint.
Ziye Ma
Download here
Topology-Aware 3D Gaussian Splatting: Leveraging Persistent Homology for Optimized Structural Integrity
39th AAAI Conference on Artificial Intelligence 2025, Oral presentation
Tianqi Shen, Shaohua Liu, Jiaqi Feng, Ziye Ma, Ning An
Download here
Absence of spurious solutions far from ground truth: A low-rank analysis with high-order losses
International Conference on Artificial Intelligence and Statistics(AISTATS) 2024, Oral presentation
Ziye Ma*, Ying Chen*, Javad Lavaei, Somayeh Sojoudi
Download here
Algorithmic Regularization in Tensor Optimization: Towards a Lifted Approach in Matrix Sensing
Neural Information Processing Systems (NeurIPS) 2023.
Ziye Ma, Javad Lavaei, Somayeh Sojoudi
Download here
Over-parametrization via Lifting for Low-rank Matrix Sensing: Conversion of Spurious Solutions to Strict Saddle Points
International Conference on Machine Learning(ICML) 2023, Oral presentation
Ziye Ma, Igor Molybog, Javad Lavaei, Somayeh Sojoudi
Download here
Semidefinite Programming versus Burer-Monteiro Factorization for Matrix Sensing
37th AAAI Conference on Artificial Intelligence 2023, Oral presentation
Baturalp Yalcin*, Ziye Ma*, Javad Lavaei, Somayeh Sojoudi
Download here
Noisy Low-rank Matrix Optimization: Geometry of Local Minima and Convergence Rate
International Conference on Artificial Intelligence and Statistics(AISTATS) 2023, Oral presentation (Top 32 notable papers).
Ziye Ma, Somayeh Sojoudi
Download here
Sharp Restricted Isometry Property Bounds for Low-rank Matrix Recovery Problems with Corrupted Measurements
36th AAAI Conference on Artificial Intelligence 2022.
Ziye Ma*, Yingjie Bi*, Javad Lavaei, Somayeh Sojoudi
Download here
Geometric Analysis of Noisy Low-rank Matrix Recovery in the Exact Parameterized and the Overparameterized Regimes
INFORMS Journal on Optimization
Ziye Ma, Yingjie Bi, Javad Lavaei, Somayeh Sojoudi
Download here
A Sequential Framework Towards an Exact SDP Verification of Neural Networks
2021 IEEE 8th International Conference on Data Science and Advanced Analytics (DSAA)
Ziye Ma, Somayeh Sojoudi
Download here
Tightened convex relaxations for neural network robustness certification
2020 59th IEEE Conference on Decision and Control (CDC)
Brendon G Anderson, Ziye Ma, Jingqi Li, Somayeh Sojoudi
Download here
Certifiably globally optimal extrinsic calibration from per-sensor egomotion
IEEE Robotics and Automation Letters
Matthew Giamou, Ziye Ma, Valentin Peretroukhin, Jonathan Kelly
Download here