CV

About me

I am a fifth-year Machine Learning Ph.D. candidate in the H. Milton Stewart School of Industrial and Systems Engineering (ISyE) at Georgia Tech and a research assistant at NSF Artificial Intelligence Research Institute for Advances in Optimization (AI4OPT). My research with Prof. Pascal Van Hentenryck focuses on developing reliable, efficient, and scalable decision-making solutions by fusing Machine Learning and Mathematical Optimization for systems at a massive scale driven by societal challenges in energy, supply chains and sustainability. The methodology can be categorized into the following two areas:

1. Machine Learning for Real-Time Optimization: Enable real-time decision making at scale by developing optimization proxies (i.e., computationally efficient ML surrogates) that approximate computationally expensive optimization models.
2. Optimization for Reliable Machine Learning: Enable reliable and robust ML models by developing scalable differentiable optimization layer to ensure models’ outputs satisfy hard constraints, designing formal verification to prove models’ exact robustness and developing conformal prediction to provide statistical guarantees on model performances.

Education

• Ph.D. in Machine Learning, Georgia Institute of Technology, Expected in May 2024
  • Minor: Operations Research
  • The H. Milton Stewart School of Industrial and Systems Engineering (ISyE)
• B.S. in Electrical Engineering, Huazhong University of Science and Technology, 2019

Professional experiences

• Jan 2021 - Present: Graduate Research Assistant in AI4OPT
  • Research: Integrating machine learning and optimization to enable reliable intelligent decision making at massive scale,
  • Service: Hosting AI4OPT student ML methodology reading group & Mentoring junior Ph.D. students and undergraduates.
• Jan 2023 - Present: Research Intern in Kinaxis
  • Developing learning-augmented optimization solver for rapid response (Kinaxis core product) in the supply chain planning
• Jan 2021 - May 2023: Research Assistant for Industrial Collaborator in MISO [5][7][8][9][10][13][14]
  • Proposed the first ML surrogate to large-scale security-constrained economic dispatch problem on RTE system with 6,708 buses in the MISO pipeline. The proposed proxies produce the optimal dispatches with relative errors 0.6% within milliseconds [5].
  • Proposed end-to-end, self-supervised ML surrogate for DC optimal power flow on up to 30,000 buses system (the largest open-sourced system) with feasibility guarantee. It achieves 5 orders of magnitude faster than Gurobi (the fastest commercial solver) with the optimality gap less than 0.5% [9].
  • Proposed confidence-aware graph neural network to accelerate solving security-constrained unit commit-ment on RTE system (a mixed-integer linear program with millions of decision variables and constraints). It generates feasible solutions with 0.77% optimality gap with 4 times speedup than Gurobi [7].
• Jun 2020 - Aug 2020: Research Intern in Ant Financial
• Aug 2019 - Present: Graduate Research Assistant
  • Georgia Institute of Technology

Publications

Pre-prints and working papers:

[14] Wenbo Chen, Mathieu Tanneau and Pascal Van Hentenryck. Optimality Verification for Optimization Proxies.

[13] Wenbo Chen, Mathieu Tanneau and Pascal Van Hentenryck. Real-Time Risk Assessment with Optimization Proxies.

[12] Ritesh Ojha*, Wenbo Chen*, Hanyu Zhang, Reem Khir, Alan Erera and Pascal Van Hentenryck. Optimization-based Learning for Load Plan Modification in Service Networks. [Paper], submitted to Transportation Science. *co-first author

[11] Wenbo Chen, Reem Khir and Pascal Van Hentenryck. Two-Stage Learning For the Flexible Job Shop Scheduling Problem. [Paper]

[10] Oliver Stover, Pranav Karve, Sankaran Mahadevan, Wenbo Chen, Haoruo Zhao, Mathieu Tanneau, Pascal Van Hentenryck. Just-In-Time Learning for Operational Risk Assessment in Power Grids. [Paper], submitted to IEEE Transactions on Power Systems.

Publications:

[9] Wenbo Chen, Mathieu Tanneau and Pascal Van Hentenryck. End-to-End Feasible Optimization Proxies for Large-Scale Economic Dispatch. IEEE Transactions on Power Systems, Sept. 2023. [Paper]

[8] Seonho Park, Wenbo Chen, Terrence W.K. Mak and Pascal Van Hentenryck. Compact Optimization Learning for AC Optimal Power Flow. IEEE Transactions on Power Systems, Sept. 2023. [Paper]

[7] Seonho Park, Wenbo Chen, Dahye Han, Mathieu Tanneau, Pascal Van Hentenryck. Confidence-Aware Graph Neural Networks for Learning Reliability Assessment Commitments. IEEE Transactions on Power Systems, Jul. 2023. [Paper]

[6] Enpeng Yuan, Wenbo Chen, Pascal Van Hentenryck. Reinforcement Learning from Optimization Proxy for Ride-Hailing Vehicle Relocation. Journal of Artificial Intelligence Research, 2022. [Paper]

[5] Wenbo Chen, Seonho Park, Mathieu Tanneau, Pascal Van Hentenryck. Learning Optimization Proxies for Large-scale Security-Constrained Economic Dispatch. PSCC-EPSR, 2022. [Paper]

[4] Haoran Sun, Wenbo Chen, Hui Li, Le Song. Improving Learning to Branch via Reinforcement Learning. Learning Meets Combinatorial Optimization Workshop, NeurIPS, 2020. [Paper]

[3] Wenbo Chen, Anni Zhou, Pan Zhou, Liang Gao, Shouling Ji, and Dapeng Oliver Wu. Privacy-Preserving Online Learning Approach for Incentive-based Demand Response in Smart Grid. IEEE System Journal, 2019, [Paper]

[2] Pan Zhou, Wenbo Chen, Shouling Ji, Hao Jiang, Li Yu and Dapeng Oliver Wu. Privacy-Preserving Online Task Allocation in Edge-Computing-Enabled Massive Crowdsensing. IEEE Internet of Things Journal, 2019. [Paper]

[1] Wenbo Chen, Pan Zhou, Shaokang Dong, Shimin Gong, Menglan Hu, Kehao Wang, and Dapeng Oliver Wu. Tree-based Contextual Learning for Online Job or Candidate Recommendation with Big Data Support in Professional Social Networks. IEEE ACCESS, 2018. [Paper]

Invited Presentations:

• INFORMS Conference, “End-to-End Feasible Optimization Proxies for Large-Scale Economic Dispatch”, October 2023, Phoenix, AZ
• NSF AI4OPF Annual Review, “End-to-End Learning and Optimization”, Jun 2023, Atlanta, GA
• IISE Conference, “Confidence-Aware Graph Neural Networks for Learning Reliability Assessment Commitments”, May 2023, New Orleans, LA
• Google Research Operations Research, “Machine Learning for Discrete Optimization and Applications in Power Systems and Supply Chains”, May 2023, Virtual
• INFORMS Conference, “Learning Optimization Proxies for Large-Scale Security-Constrained Economic Dispatch”, October 2022, Indianapolis, IN
• ISyE Student Seminar, “Learning Optimization Proxies for Large-Scale Security-Constrained Economic Dispatch”, April 2022, Georgia Tech