Artificial intelligence as a catalyst for combustion science and engineering

Combustion and energy conversion play critical roles in all facets of environmental and technological applications, including the utilization of sustainable energy sources for power generation and propulsion, the reduction of pollutant emissions from combustion, or the mitigation of harm from wildfire hazards. Computational and mathematical tools have long been crucial in combustion research in the form of highfidelity simulations, dynamical-system modeling, and data analytics. With the advent of data-driven methods, machine learning (ML) and artificial intelligence (AI) offer numerous opportunities for predictive modeling, improving existing research methods, and extracting new knowledge from data. In this article, we discuss recent progress on how ML and AI can impact the field of combustion and energy conversion, and discuss the need for domain knowledge for successful ML applications in combustion. Specifically, combustion ML learns from data extracted from large-scale simulations, high-resolution experiments, and sensors, which can introduce challenges tied to dimensionality, interpretability, sparsity, modality, and scarcity. The collective knowledge from these advancements equip combustion researchers and engineers with the ability to adapt to emerging developments in ML foundation models and AI agents, which have begun to offer greater automation across different combustion domains. To this end, we assess opportunities and challenges provided by stateof-the-art ML foundation models, and discuss emerging areas for adapting these new technologies towards solving pressing challenges within sustainable combustion.