Explore a 29-minute conference talk on fast semidefinite programming for differentiable combinatorial optimization. Delivered by Zico Kolter from Carnegie Mellon University at the Deep Learning and Combinatorial Optimization 2021 event, hosted by the Institute for Pure and Applied Mathematics at UCLA. Dive into advanced techniques for solving complex optimization problems, combining elements of deep learning and combinatorial methods. Gain insights into cutting-edge research that bridges the gap between continuous and discrete optimization, potentially revolutionizing approaches to challenging computational tasks.
Fast Semidefinite Programming for Differentiable Combinatorial Optimization
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Syllabus
Zico Kolter: "Fast semidefinite programming for (differentiable) combinatorial optimization"
Taught by
Institute for Pure & Applied Mathematics (IPAM)
Related Courses
-
Differentiable Programming for End-to-end Optimization of Experiments - Lecture 5
-
Combining Reinforcement Learning and Constraint Programming for Combinatorial Optimization
-
Deep Learning for Combinatorial Optimization - Count Your Flops & Make Your Flops Count
-
Decision-Focused Learning - Integrating Downstream Combinatorics in ML
-
A Framework for Differentiable Discovery of Graph Algorithms
-
Differentiable Programming for End-to-end Optimization of Experiments - Lecture 6
