Our research interest includes modeling, optimization techniques and theories, and deep learning architectures for high dimensional data analysis. Current ongoing projects are
- Deep learning architectures inspired by optimization method: An integration of variational method and deep neural network (DNN) approach for data analysis;
- Variational DNN for medical image processing;
- Develop randomized incremental primal dual algorithms for decentralized consensus problems to reduce communication cost for fully connected networks;
- Develop accelerated level bundle methods for functional constrained convex optimization.
- Develop sparsity-induced stochastic ADMM algorithm for non-convex stochastic optimization to achieve better iteration complexity and sample complexity (joint with Dr. Liu, ISE, UF)
Leading figure shows Prof. Yunmei Chen and her students.