Dear Colleagues,
You are cordially invited to join 2023 Dartmouth Summer Data Science and AI Webinar Series - Week 4. Please take a look at the details of the event below.
Title: From Poly-disentanglement to AccoMontage: Modular Concept Learning for Music Arrangement
Date/Time: July 27, Thursday, 11:00 AM – 12:00 PM EST
Abstract:
One day, Mr. Ran Wang, a professional pianist friend of Gus visited the Music X Lab at NYU Shanghai and showcased his amazing techniques of improvisational accompaniment. Gus, as a Music AI researcher as well as a piano beginner, was curious how that is even possible and asked Ran for his secrete. Ran said it's easy. First, given a lead melody, I will search for songs (piano pieces) I have played that could possibly serve as the "raw materials" for the accompaniment. Second, I deconstruct the raw materials and mainly retain the piano texture. Finally, I re-construct the materials with a new harmonization I composed on the fly for the lead melody. And quite simultaneously, Gus translates it into a research plan: (1) disentangle the piano representation, (2) style transfer using new harmony, and (3) global planning/searching on representation space. These efforts lead to a series of models, including poly-disentanglement and AccoMontage I, II, and III, each one learning a modular concept and solving a more complex task based on the previous one.
Bio of the speaker: Gus Xia is a faculty member at MBZUAI. He leads the Music X Lab at NYU Shanghai, building various intelligent systems that help people better compose, perform, and learn music. Some representative works include music style transfer, interpretable music concept learning, self-supervised MIR, and music robotics.
This webinar will be offered online via Zoom. Please register to receive the Zoom link (one day before the webinar).
Registration link: libcal.dartmouth.edu/calendar/itc/2023DSAIW4
Look forward to seeing you at the webinar!
------------------------------
Jianjun Hua
Dartmouth College
------------------------------