Ph.D. Research Proposal: Mingju Liu

Tuesday, February 3, 2026
1:30 p.m.
AVW 1146

ANNOUNCEMENT: Ph.D. Research Proposal Exam

 

Name: Mingju Liu

 

Committee:

Professor Cunxi Yu (Chair)

Professor Kaiqing Zhang

Professor Gang Qu

 

Date/time: Tuesday, February 3, 2026 at 1:30pm

 

Location: AVW 1146

 

Title: Sampling-Opt: Bridging Differentiable Learning and Combinatorial Optimization for Next-Generation EDA

 

Abstract: 

Combinatorial optimization lies at the heart of Electronic Design Automation (EDA), spanning tasks from high-level scheduling to logic synthesis and verification. Traditional methods face a critical trade-off: exact solvers (such as ILP and SAT solvers) guarantee optimality but suffer from poor scalability due to exponential time complexity, while heuristic-based methods offer speed at the cost of sub-optimal results.

My research proposes a paradigm called "Sampling-Opt" to advance the resolution of classic combinatorial optimization problems in EDA. This approach leverages various sampling-based techniques—specifically the Differentiable Gumbel-Softmax trick and Reinforcement Learning (RL)—to navigate massive search spaces in specific fields efficiently. By treating discrete optimization decisions as differentiable sampling steps, we can rapidly identify high-quality solutions or significantly reduced search spaces.

Furthermore, I propose a hybrid methodology that employs traditional solvers to refine the solutions obtained from sampling-based methods. This approach aims to combine the scalability of learning-based sampling with the optimality of classic exact solvers. Beyond theoretical contributions, I will demonstrate real-world applications capable of being extended to industrial-level tools.

Audience: Graduate  Faculty 

remind we with google calendar

 

January 2026

SU MO TU WE TH FR SA
28 29 30 31 1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30 31
1 2 3 4 5 6 7
Submit an Event