Challenges and Opportunities in Machine Learning for Light-Emitting Polymers

Light-emitting polymers (LEPs) combine the luminescent properties of organic emitters with the structural versatility of polymers, supporting applications in solid-state display, chemical sensing, and bioimaging, owning to the efficient tuning of their performance across multiple scales, from monomer units and chain sequence to solid-state packing and solution processing. Recent strategies have expanded emission color space, improved quantum yields, and simplified design rules, evolving from traditional π-conjugated systems to mechanisms driven by aggregation and charge transfer. Yet this multiscale flexibility also creates a vast and complex design space, where the interplay of monomer choice, polymer architecture, and processing methods makes it impossible to exhaustively map their structure–property relationships by empirical means. In this perspective, we review the development of recent design strategies in LEPs, highlighting the key experimental challenges they reveal, and discuss how data-driven approaches, particularly machine learning, can help navigate this complexity and accelerate the discovery and optimization of next-generation LEPs.