2026-08-13 –, Room 6
Decision-Focused Learning (DFL) is a field at the intersection of machine learning and combinatorial optimization. It integrates prediction with combinatorial decision-making by embedding optimization algorithms directly into machine learning pipelines. This talk presents the JuliaDecisionFocusedLearning ecosystem, focusing on DecisionFocusedLearningBenchmarks.jl and DecisionFocusedLearningAlgorithms.jl, two new packages that provide a high-level and generic interface for using state-of-the-art DFL methods.
This talk provides a practical introduction to Decision-Focused Learning (DFL) in Julia. For a recent survey of the field, see https://arxiv.org/abs/2601.10583.
Rather than focusing on theoretical details, the goal is to use a small example problem to showcase the capabilities of the JuliaDecisionFocusedLearning ecosystem.
JuliaDecisionFocusedLearning started with InferOpt.jl, which provides the core building blocks for constructing differentiable combinatorial optimization layers and associated loss functions. We introduce two new higher-level packages:
- DecisionFocusedLearningBenchmarks.jl
- Provides a growing collection of benchmark combinatorial decision problems.
- Includes all necessary components to build and train a DFL policy for each problem.
- Offers a general interface for defining custom problems.
- Facilitates reproducibility and experimentation by allowing the same algorithm to be applied across multiple benchmarks with minimal changes.
- DecisionFocusedLearningAlgorithms.jl
- Implements generic versions of state-of-the-art DFL training algorithms.
- Provides compatibility with problems defined through the benchmarks interface.
- While some classical DFL approaches can be implemented directly using the lower-level tools provided by InferOpt.jl, recent state-of-the-art methods have become increasingly complex. This package introduces higher-level implementations to make these advanced algorithms accessible without requiring users to master the technical details.
The goal of the talk is to demonstrate how users can leverage these packages to define a problem, select a training strategy, and train a policy for their own problem in a few lines of code, without necessarily needing deep expertise in DFL.
Research engineer in combinatorial optimization and machine learning. Member of JuliaDecisionFocusedLearning.