First Workshop on Reusable Learning for Solving Engineering Systems

AI4Engineers| Imperfect Physics| Perfect Physics| Sparse Data| Sustainable Learning

Aim: Learning based approaches are increasingly shaping scientific and engineering discovery. Thus far, much of the work in this area has focused on improving benchmark performance under idealized physical assumptions or developing methods tailored to specific domains with abundant data and compute. In this workshop, we wish to focus on a complementary and less established direction: solving real-world engineering problems that lie between perfect physical models and imperfect or incomplete physical systems through reusable learning. Our goal is to promote a sustainable and accessible paradigm for learning physical and engineering systems, where models are not only accurate, but also reusable, efficient, and reliable. In particular, we emphasize approaches that can be developed under practical constraints such as sparse data, limited computational resources, and small-lab settings, while enabling reuse and transfer of trained models across tasks, domains, and operating conditions. By prioritizing sustainability and reusability, this workshop aims to lower adoption barriers and support a more inclusive ecosystem in which learning-based engineering tools benefit the broader community.

Topics of interest:

Physics-informed Machine Learning for Complex Engineering Systems
Data-Efficient Learning in Engineering Applications
Hybrid Learning and Modeling for Engineering Problems
Generalization and Transfer across Engineering Scenarios
Reusable and Modular Learning Systems for Engineering
Reusability of Foundation models, agents based learning systems for Engineering
Learning for Imperfect Physics (Hybrid + Model-Mismatch Systems)
Engineering Principles for General ML

Submission Tracks

All submissions must use the official ICML 2026 workshop template and follow the ICML formatting guidelines. There are two tracks, i) original research track which allows the submission of unpublished work, ii) research communication track which allows only the presentation of published work without being included in proceedings. The purpose of the research communication track is to foster discussion among top-tier researchers.

Format of papers

Short Papers (4–6 pages, excluding references):
Focused contributions including early-stage ideas, methodological advances, ablation studies, or application-driven insights.

Extended Abstracts (2–3 pages, excluding references):
Preliminary results, position papers, benchmarks, negative results, or work-in-progress that may benefit from discussion and feedback.