DS-GA 1011 (Fall 2021) - Lecture 13 - AI for Music and Music for AI by Keunwoo Choi |
|
Presenter: Keunwoo Choi (https://keunwoochoi.github.io)
Slides: https://www.slideshare.net/KeunwooChoi/all-you-need-is-ai-and-music-by-keunwoo-choi ---------- [0/5] ---------- 0:06 ๐ Introduction by Kyunghyun Cho 1:00 Keunwoo's starting / Abstract / Content ---------- [1/6] ---------- 3:31 ๐ Starting the main content 3:40 What is Music AI? * An On-line algorithm for real-time accompaniment, Roger Dannenberg, 1984, ICMC, https://www.cs.cmu.edu/~rbd/papers/icmc84accomp.pdf 7:14 2-by-2 Categorization of Music AI ---------- [2/6] ---------- 9:27 ๐ Music AI - 1) Audio Synthesis * Synthesizer / Three components of sound / Autoencoder * DDSP: Differentiable Digital Signal Processing, Jesse Engel et al., 2019, ICLR, https://magenta.tensorflow.org/ddsp * Tone Transfer https://sites.research.google/tonetransfer * AI Sax Demo by Hanoi https://twitter.com/HanoiHantrakul/status/1258803013948342272 * Karasinger:Score-free singing voice synthesis with VQ-VAE using mel-spectrograms, 2021, Chien-Feng Liao, https://jerrygood0703.github.io/KaraSinger/ * RAVE: A variational autoencoder for fast and high-quality neural audio synthesis, Antoine Caillon and Philippe Esling, 2021, https://arxiv.org/abs/2111.05011 ---------- [3/6] ---------- 31:08 ๐ Music AI - 2) Creation * Text-based LSTM networks for Automatic Music Composition, Keunwoo Choi et al., 2016, CSMC, https://arxiv.org/abs/1604.05358# * Music Transformer: Generating Music with Long-Term Structure, Cheng-Zhi Anna Huang et al., 2019, ICLR, https://magenta.tensorflow.org/music-transformer * MIDI-VAE: Modeling Dynamics and Instrumentation of Music with Applications to Style Transfer, Gino Brunner et al., 2018, ISMIR, https://arxiv.org/abs/1809.07600 * Pop Music Transformer: Beat-based Modeling and Generation of Expressive Pop Piano Compositions, Yu-Siang Huang and Yi-Hsuan Yang, 2020, ACM MM, https://arxiv.org/abs/2002.00212 ---------- [4/6] ---------- 48:27 ๐ Music AI - 3) Audio Signal Processing * Source Separation demo, Qiuqiang Kong, 2020, https://www.youtube.com/watch?v=WH4m5HYzHsg * Steerable discovery of neural audio effects, Christian Steinmetz and Joshua Reiss, 2021, NeurIPS Creativity workshop, https://csteinmetz1.github.io/steerable-nafx/ ---------- [5/6] ---------- 1:01:32 ๐ Music AI - 4) Analysis 1:03:06 ๐๐ Timbre Understanding * MFCC, Convnets * Automatic tagging using deep convolutional neural networks, Keunwoo Choi et al., 2016 ISMIR, https://arxiv.org/abs/1606.00298 * Music Classification Tutorial (https://music-classification.github.io/tutorial/) 1:17:47 ๐๐ Note-level Understanding * Onsets and Frames: Dual-Objective Piano Transcription, Curtis Hawthorne et al., 2018 ISMIR, https://arxiv.org/abs/1710.11153 * Polyphonic Piano Transcription Using Autoregressive Multi-State Note Model, Taegyun Kwon et al., 2020 ISMIR, https://arxiv.org/abs/2010.01104 * High-resolution Piano Transcription with Pedals by Regressing Onset and Offset Times, Qiuqiang Kong et al., 2020, https://arxiv.org/abs/2010.01815 * The MAESTRO Dataset and Wave2Midi2Wave, Curtis Hawthorne et al., ICLR 2019, https://magenta.tensorflow.org/maestro-wave2midi2wave * Deep Unsupervised Drum Transcription, Keunwoo Choi and Kyunghyun Cho, ISMIR 2019, https://arxiv.org/abs/1906.03697 1:33:44 ๐๐ Lyric Understanding ---------- [6/6] ---------- 1:35:08 ๐ Conclusion * ISMIR2021 Lab Showcase https://ismir2021.ismir.net/labshowcase/ |