Title: Learning-augmented algorithms

Abstract: Learning-augmented algorithms, also known as algorithms with predictions, are a rapidly growing research area that integrates traditional algorithm design with machine-learned predictions. By incorporating potentially imperfect predictions, the objective of these algorithms is to improve upon worst-case performance guarantees, while remaining robust against arbitrary prediction errors. This paradigm has been successfully applied to the design of data structures such as binary search trees, dictionaries, skip lists, and priority queues, as well as to algorithms in distributed systems, including scheduling, load balancing, and network and graph algorithms. This talk aims to formally introduce learning-augmented algorithms, provide an overview of related key advancements in data structures and distributed systems, and explore the main challenges and future research directions in the field.