An MIT profile published Oct. 28, 2026 describes the machine-listening research of Paris Smaragdis, a professor in MIT's Music and Theatre Arts program who holds a shared position with the Department of Electrical Engineering and Computer Science. He joined MIT in 2025, according to the profile.

Smaragdis's group works on machine learning for sound source separation, devices that augment human audio surroundings, and tools to help people understand, analyze and create music, the profile says. It describes recent exploration of signal-processing algorithms that do not depend on fixed sampling rates, frame sizes or context windows, and of incorporating large language models into speech recognition and audio enhancement. Smaragdis says AI is powerful but that an ideal way to use it for sound and music has not been found, the profile reports; it gives no experimental results or clinical claims.

The profile also recounts earlier work, including neural networks for separating sounds from audio mixtures that drew the attention of the team behind the 2021 documentary series "The Beatles: Get Back." It says his contributions helped that team win a 2022 Emmy for sound editing.