Overview

Bridging Machine Learning (ML) \& Combinatorial Optimization (CO) -- from ML for CO to CO for ML. Advancing foundations, scalability, and real-world AI-driven decision making.

Machine learning (ML) and combinatorial optimization (CO) are two foundational paradigms for solving large-scale real-world problems. ML maps data into high-dimensional continuous parameter spaces and optimizes models through gradient-based, data-driven learning, enabling breakthroughs in image recognition, language processing, and robotics. In contrast, CO focuses on complex discrete decision-making, leveraging advanced search, sampling, and branch-and-cut algorithms to address challenges in scheduling, planning, routing, hardware/chip design, drug discovery, etc.

In recent years, the synergy between ML and CO has expanded exponentially. ML has increasingly accelerated CO solvers through automated heuristic discovery, learned search guidance, especially for nonconvex, uncertain, and large-scale problems that are evasive to off-the-shelf human solvers. Reciprocally, CO provides the fundamental reasoning and constrained search frameworks integral to modern ML systems, enhancing their expressiveness, efficiency, and reliability from training to inference. Looking ahead, CO serves a pivotal domain distinguished by different inductive biases (say, from vision and language), such as decision-making landscape representation and reasoning, motivating next-generation AI systems for countless applications. Together, these trends signal a deep coevolution of ML and CO that is reshaping the future of AI-driven decision-making.

The workshop will feature invited speakers across two complementary tracks: (a) ML for CO Track,, focusing on ML-driven methods for CO, and (b) CO for ML Track, showcasing how combinatorial techniques, particularly discrete searching/sampling/reasoning, can enhance ML. By bringing together these perspectives, the workshop aims to capture this emerging direction, foster cross-community dialogue, and position the ML-CO interface as a timely research agenda for the ICML community.

Scope and Goals

This workshop explores the growing intersection of machine learning (ML) and combinatorial optimization (CO), with a focus on both ML4CO and CO4ML. We welcome contributions across a wide range of topics within ML4CO, including but not limited to Differentiable CO solver, Neural CO solver/Generative CO solver, and Automatic CO Algorithm Discovery. On the CO4ML side, we invite works that leverage combinatorial frameworks to address the inherent discrete challenges within structured machine learning tasks, such as Combinatorial Matching&Structural Alignment and Decoding/Search Algorithm for LLMs.

By bringing together these communities, we aim to foster new collaborations and inspire innovative research directions at the intersection of ML and CO. By fostering shared understanding and encouraging contributions to common infrastructure—such as benchmarks, scalable tools, and system-level integrations—we hope to accelerate mutual progress and strengthen the foundation for future breakthroughs.

Speakers & Panelists

Cathy Wu

MIT

Cathy Wu is an associate professor at MIT Department of Civil and Environmental Engineering (CEE) and the Institute for Data, Systems, & Society (IDSS). Her research focuses on using machine learning to tackle the challenging optimization and control problems that are prevalent in transportation systems.

Topic: Hybridizing machine learning model-based methods for optimization.

Changhyun Kwon

KAIST

Changhyun Kwon is Associate Professor in Industrial and Systems Engineering at KAIST. His current focus is to improve the efficiency of heuristic and exact algorithms using machine-learning approaches to solve large-scale vehicle routing problems and mobility service operations problems.

Topic: Learning-Based Separation Algorithms for Cutting Planes in CO.

Alexander Novikov

Google DeepMind

Alexander Novikov is a Research Scientist at Google DeepMind. His work focuses on foundational and applied problems in machine learning, with a particular emphasis on algorithmic discovery. Recently, he has contributed to advances in automated algorithm design for challenging mathematical problems using large language models.

Topic: AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery.

Junchi Yan

Shanghai Jiao Tong University

Junchi Yan is a Full Professor in the Department of Computer Science and Engineering at Shanghai Jiao Tong University. His research focuses on machine learning for real-world decision-making and intelligent systems. He previously spent 10 years in industry research and advising roles, including at IBM Research and Amazon.

Topic: Structure-Aware Learning and Optimization Landscapes in Foundation Models.

Cunxi Yu

NVIDIA

Cunxi Yu is a Senior Research Scientist at NVIDIA Research. She develops AI solutions for decision-making, focusing on combining machine learning and deep reinforcement learning with data-driven optimization and socially aware algorithms.

Topic: Application of GNNs and LLMs in Solving Real-World CO Problems.

Maximilian Schiffer

Technical University of Munich

Maximilian Schiffer is Associate Professor of Business Analytics & Intelligent Systems at TUM School of Management, Technical University of Munich (TUM). His research focuses on developing operations research and prescriptive analytics methods to solve central societal problems, especially in the field of mobility and transportation.

Topic: Combinatorial Optimization Augmented Machine Learning.

Panelists

Jinkyoo Park

KAIST

Jinkyoo Park is an Associate Professor in Industrial and Systems Engineering at KAIST, and also the founder and CEO of Omelet, a startup developing AI-based CO solvers. His research focuses on AI-based CO, reinforcement learning, and ML for industrial applications, as well as data-driven decision-making.

Hyeonah Kim

Mila & Université de Montréal

Hyeonah Kim is a postdoctoral researcher at Mila and Université de Montréal. Her research interests lie in scientific discovery with deep learning, with a particular interest in GFlowNets, active learning, and sample-efficient training.

Call For Papers

For detailed submission guidelines and more information, please refer to the call for papers.

Workshop Schedule

Organizers

Sirui Li

Senior Research Engineer, Microsoft Research

Yining Ma

Postdoctoral Associate, MIT

Zhiguang Cao

Assistant Professor, SMU

Eli Meirom

Senior Research Scientist, NVIDIA

Shengyu Feng

PhD Student, CMU

Yoonju Sim

PhD Student, KAIST

Community Building

Collaboration across machine learning, operations research, and diverse ML applications is essential for the success of ML and CO. This workshop will serve as a platform to spark dialogue, inspire novel solutions, and drive future advances. We aim to build lasting connections among researchers, practitioners, and industry professionals. Designed as the start of an ongoing workshop series, we will foster a growing, supportive community with long-term engagement.

We have launched the AI4CO Community which includes an active Slack channel that already hosts 400+ members and is open to all interested in ongoing collaboration. Join our community to stay connected with the latest developments in the intersection of AI and combinatorial optimization.

Venue