אודות פרק זה
In this episode, Prof. Hung-yi Lee, Professor at National Taiwan University, interviews Dr. Jinyu Li, Partner Applied Science Manager at Microsoft. Their conversation traces the evolution of speech translation from traditional cascaded pipelines to today's end-to-end, large language model (LLM)-powered systems.
Jinyu Li
Dr. Jinyu Li is Partner Applied Science Manager at Microsoft, where he leads a science team advancing speech, translation, and language technologies. He is an IEEE Fellow and an ISCA Fellow for contributions to deep-learning-based speech technology innovation, adoption, and commercialization. He has served on the IEEE Speech and Language Processing Technical Committee and as Vice Chair starting in 2026. He previously served as Associate Editor of IEEE/ACM Transactions on Audio, Speech, and Language Processing and is a Distinguished Industry Speaker for the IEEE Signal Processing Society. He has also been awarded the IEEE SPS Best Paper Award (2025), the APSIPA Industrial Distinguished Leader Award (2021), and the APSIPA Sadaoki Furui Prize Paper Award (2023).
In this episode, Dr. Li discusses the evolution of AI-powered speech translation, from conventional cascaded systems to modern end-to-end architectures, while highlighting the role of large language models in enabling real-time multilingual communication. He also explores the latest applications of speech translation and examines the technical challenges that remain.
