About me

KuoLing Huang

I am a Principal Scientist at Anthem. I was previously working as a Research Associate in the department of IE/MS at Northwestern University.

My areas of research interest are

• Computational optimization and software development.
• Machine learning, AI, NLP in healthcare.

Personal Information

• E-mail
• Personal website
• Colin optimization software
• Google scholar
• Linkedin
• Resume

Biography

KuoLing Huang is currently a Principal Scientist at Anthem. He holds a Ph.D. degree of IE/MS from Northwestern University, and a M.S. degree of Industrial Management from National Taiwan University of Science and Technology.

KuoLing’s research interests include studies efficient algorithms for solving large-scale optimization problems. In particular, he has studied and implemented a homogeneous interior point method for linear, conic, and general convex optimization, as well as a Benders decomposition method for two-stage stochastic and distributionally-robust optimization with discrete recourse. He has also studied deterministic algorithms and randomized search heuristics for solving discrete optimization problems. He has implemented and experimented a novel branching algorithm for mixed-integer convex optimization problems, and has developed different heuristics based on geometric random walks for generating high quality feasible solution for the same problems.

Kuo-Ling has also worked on machine scheduling, which is considered as one of the most important applications in discrete optimization. Specifically, he has studied using deterministic method (mixed-integer programming approach) as well as randomized search heuristics (meta-heuristic approach) for solving challenging scheduling problems.

In additional to the area of computational optimization, KuoLing is also interested in machine learning. He is currently managing Anthem's proprietary auto-machine learning software. This package includes an optimized auto-machine learning pipeline that allows data scientists/analysts creating effective predictive models efficiently, visualizing model training results clearly, performing a what-if analysis for feature explanations effectively, and deploying the models into production line smoothly. This package has been used to deliver hundreds of predictive models to date.