Machine Learning and Data Science Research
Data plays a critical role in all areas of IEOR, from theoretical developments in optimization and stochastics to applications in automation, logistics, health care, energy, finance, and other areas. Much of the recent interest in data science and machine learning has been spurred by the growing ability to apply vast computational power to large scale datasets in nearly every application domain. Faculty and students in the UC Berkeley IEOR department are engaged in cutting edge and interdisciplinary research in ML/DS, including topics like developing scalable and memory-efficient learning algorithms, integrating prediction and optimization models, sparse learning models, addressing fairness concerns, reinforcement learning and control, clustering and learning with network data, as well as applications of ML/DS to various domains.
Faculty
Selected Publications
On Efficient and Scalable Computation of the Nonparametric Maximum Likelihood Estimator in Mixture Models
Zhang, Y., Cui, Y., Sen, B., & Toh, K. (2022). On Efficient and Scalable Computation of the Nonparametric Maximum Likelihood Estimator in Mixture Models. ArXiv. /abs/2208.07514
A New Perspective on Boosting in Linear Regression via Subgradient Optimization and Relatives
Freund, Robert & Grigas, Paul & Mazumder, Rahul. (2015). A New Perspective on Boosting in Linear Regression via Subgradient Optimization and Relatives. The Annals of Statistics. 45. 10.1214/16-AOS1505.
Smart “Predict, then Optimize”
Elmachtoub, Adam & Grigas, Paul. (2017). Smart “Predict, then Optimize”. Management Science. 68. 10.1287/mnsc.2020.3922.
The Non-Markovian Nature of Nested Logit Choice
Li, Selena & Udwani, Rajan. (2022). The Non-Markovian Nature of Nested Logit Choice. SSRN Electronic Journal. 10.2139/ssrn.3420257.
Learning to Simulate Sequentially Generated Data via Neural Networks and Wasserstein Training
Zhu, Tingyu & Liu, Haoyu & Zheng, Zeyu. (2023). Learning to Simulate Sequentially Generated Data via Neural Networks and Wasserstein Training. ACM Transactions on Modeling and Computer Simulation. 10.1145/3583070.
Mean-Field Controls with Q-Learning for Cooperative MARL: Convergence and Complexity Analysis
Logarithmic regret for episodic continuous-time linear-quadratic reinforcement learning over a finite-time horizon
M. Basei, X. Guo, A. Hu, Y. Zhang, “Logarithmic regret for episodic continuous-time linear-quadratic reinforcement learning over a finite-time horizon”. Journal of Machine Learning Research, 23 (178), 1-34
Conic Optimization for Quadratic Regression Under Sparse Noise
Molybog, Ramtin Madani, and Javad Lavaei. Conic Optimization for Quadratic Regression Under Sparse Noise. Journal of Machine Learning Research. https://www.jmlr.org/papers/v21/18-881.html.