At the Audio Information Research (AIR) Lab, we conduct research in the emerging field of computer audition: designing computational systems that are able to analyze and understand sounds including music, speech, and environmental sounds.
We address fundamental issues such as parsing polyphonic auditory scenes (the cocktail party effect), as well as designing novel applications such as sound retrieval and music information retrieval. We also combine sound analysis with the analysis of other signal modalities such as text and video towards multi-modal scene analysis. Various projects that we have been working on include audio source separation, automatic music transcription, audio-score alignment, speech enhancement, speech diarization and emotion recognition, sound retrieval, sound event detection, and audio-visual scene understanding.
We are looking for highly motivated students to join the AIR lab. Students are expected to have a solid background in mathematics, programming, and academic writing. Experience in music activities is a plus. Most importantly, students should be passionate about research in computer audition, music information retrieval, and multimodal learning.
Prospective Ph.D. students: please apply to the ECE Ph.D. program or the CS Ph.D. program and mention Prof. Zhiyao Duan in your application.
UR master's and undergraduate students: if you want to do a project or thesis in the AIR lab, please email Prof. Duan or stop by his office, Room 720, Computer Studies Building.
Our work is funded by the National Science Foundation under the following grants:
Our work is also funded by the National Institutes of Health (NIH), the National Institute of Justice (NIJ), the Intelligence Advanced Research Projects Activity (IARPA), the National Artificial Intelligence Research Resource (NAIRR) Pilot, the New York State Center of Excellence (CoE) in Data Science, University of Rochester internal awards on AR/VR, health analytics, data science, and integrated research computing, as well as gifts from Adobe, ByteDance, Kwai, IngenID, Meta, and Microsoft. We are very grateful for their support!