{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/learning-features-of-music-from-scratch","title":"Learning Features of Music from Scratch","arxiv_id":"1611.09827","date":"2016-11-29","proceeding":null,"authors":["John Thickstun","Zaid Harchaoui","Sham Kakade"],"abstract":"This paper introduces a new large-scale music dataset, MusicNet, to serve as\na source of supervision and evaluation of machine learning methods for music\nresearch. MusicNet consists of hundreds of freely-licensed classical music\nrecordings by 10 composers, written for 11 instruments, together with\ninstrument/note annotations resulting in over 1 million temporal labels on 34\nhours of chamber music performances under various studio and microphone\nconditions.\n  The paper defines a multi-label classification task to predict notes in\nmusical recordings, along with an evaluation protocol, and benchmarks several\nmachine learning architectures for this task: i) learning from spectrogram\nfeatures; ii) end-to-end learning with a neural net; iii) end-to-end learning\nwith a convolutional neural net. These experiments show that end-to-end models\ntrained for note prediction learn frequency selective filters as a low-level\nrepresentation of audio.","url_abs":"http://arxiv.org/abs/1611.09827v2","url_pdf":"http://arxiv.org/pdf/1611.09827v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"learning-features-of-music-from-scratch","repo_url":"https://github.com/benadar293/benadar293.github.io","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"learning-features-of-music-from-scratch","repo_url":"https://github.com/jthickstun/thickstun2017learning","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"multi-label-classification-2","task_name":"MUlTI-LABEL-ClASSIFICATION"},{"task_slug":"multi-label-classification","task_name":"Multi-Label Classification"},{"task_slug":"music-transcription","task_name":"Music Transcription"}],"methods":[],"datasets_introduced":[{"slug":"musicnet","name":"MusicNet","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/music-transcription-on-musicnet","task":"Music Transcription","dataset":"MusicNet","model":"CNN (64 stride)","rank_in_archive_order":6,"of":6,"metrics":{"APS":"67.8"},"uses_additional_data":false}],"syntology":{"syntology_url":"https://syntology.ai/paper/1611.09827","atlas_url":"https://app.syntology.ai/?focus=1611.09827","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}