报告题目：Lower-order regularization method for group sparse optimization with applications
胡耀华博士，深圳大学数学与统计学院副传授，硕士生导师，本科和硕士毕业于浙江大学，博士毕业于香港理工大学，从事最优化理论，算法和应用方面的研究工作。目前在最优化领域的权威期刊SIAM Journal on Optimization，European Journal of Operational Research，Journal of Global Optimization及Numerical Algorithms，Journal of Machine Learning Research和Inverse Problems等期刊上发表了多篇学术论文。
The lower-order regularization problem has been widely studied for finding sparse solutions of linear inverse problems and gained successful applications in various mathematics and applied science fields. In this talk, we will present the lower-order regularization method for (group) sparse optimization problem in three aspects: theory, algorithm and application. In the theoretical aspect, by introducing a notion of restricted eigenvalue condition, we will establish an oracle property and a global recovery bound for the lower-order regularization problem. In the algorithmic aspect, we will apply the well-known proximal gradient method to solve the lower-order regularization problem, and establish its linear convergence rate under a simple assumption. In the aspect of application, we apply the lower-order group sparse regularization method to solve two important problems in systems biology: gene transcriptional regulation and cell fate conversion.