央音要闻

学术前沿 ∣ 音乐人工智能系系列讲座——“基于小规模训练数据的音频分类”

  • 作者:供稿:音乐人工智能与音乐信息科技系  
  • 来源:
  • 发布日期:2022-04-18 15:39:00

  

  

  主讲嘉宾:ALEXANDER LERCH教授

  主持人:李小兵教授

  时间:2022年4月18日 20:00-21:30

  题目:基于小规模训练数据的音频分类

  主办:音乐人工智能与音乐信息科技系

  入场方式

  ZOOM会议号:978 438 6825

  入会密码:8888

  Title: Audio classification with insufficient data

  基于小规模训练数据的音频分类

  Topic:

  After introducing the general goals of music and audio classification, I will outline common challenges with respect to data availability impeding progress in different audio classification tasks. I will present three different recent approaches to address these challenges: (i) a semi-supervised approach, (ii) a self-supervised representation learning approach, and (iii) an approach referred to as "reprogramming." I will present preliminary results for all of them and discuss advantages and disadvantages of the various approaches.

  本次报告将首先介绍音乐与音频分类任务的通用目标,之后将通过比较半监督(semi-supervised)、自监督表示学习(self-supervised representation)和重编程(reprogramming)等方法的优势与不足,结合初步的实验结果,重点论述数据的可获取性对于各项音频分类任务带来的挑战。

  Biography:

  Alexander Lerch is Associate Professor and Director of Graduate Studies at the School of Music, Georgia Institute of Technology,Co-Chair of ISMIR 2021. He received his ``Diplom-Ingenieur'' (EE) and his PhD (Audio Communications) from TU Berlin.

  Lerch teaches computers to listen to and comprehend music. His research positions him at the intersection of signal processing, machine learning, and music and creates artificially intelligent software for music generation, production, and analysis.

  Lerch authored more than 50 peer-reviewed journal and conference papers, as well as the text book "An Introduction to Audio Content Analysis" (IEEE/Wiley 2012).

  Before he joined Georgia Tech, Lerch was Co-Founder and Head of Research at his company zplane.development, an industry leader in music technology licensing. The technologies he worked on at zplane include algorithms such as time-stretching and automatic key detection. zplane technologies are nowadays used by millions of musicians and producers world-wide.

  Alexander Lerch 现任佐治亚理工学院音乐学院副教授兼研究生院主任,ISMIR大会主席。他于柏林工业大学获得了电子工程硕士和音频通信博士学位。Lerch教授的研究集中在信号处理、机器学习和音乐的交叉结合方面,训练计算机聆听并理解音乐,并开发了一系列可用于音乐生成、制作和分析的人工智能软件。Lerch教授撰写了 50 余篇经过同行评审的期刊和会议论文,以及教材《音频内容分析导论》(IEEE/Wiley 2012)。Lerch教授曾是音乐技术领域的领头企业Z-plane公司的联合创始人兼研发主管,在任期间参与研发了音频时间伸缩与自动音高识别算法等技术成果。Z-plane在全世界范围内拥有包含音乐家和制作人在内的百万级别的用户量。

  供稿:音乐人工智能与音乐信息科技系

  文:张渊、周子雅

 
 
相关附件:
相关链接:

? Copyright www.ccom.edu.cn All Rights Reserved

京公网安备110402430057号

京ICP备05064625号

央音要闻

学术前沿 ∣ 音乐人工智能系系列讲座——“基于小规模训练数据的音频分类”

作者:供稿:音乐人工智能与音乐信息科技系来源:更新日期:2022-04-18 15:41:52发布日期:2022-04-18 15:39:00本栏目内容由党委宣传部负责维护

  

  

  主讲嘉宾:ALEXANDER LERCH教授

  主持人:李小兵教授

  时间:2022年4月18日 20:00-21:30

  题目:基于小规模训练数据的音频分类

  主办:音乐人工智能与音乐信息科技系

  入场方式

  ZOOM会议号:978 438 6825

  入会密码:8888

  Title: Audio classification with insufficient data

  基于小规模训练数据的音频分类

  Topic:

  After introducing the general goals of music and audio classification, I will outline common challenges with respect to data availability impeding progress in different audio classification tasks. I will present three different recent approaches to address these challenges: (i) a semi-supervised approach, (ii) a self-supervised representation learning approach, and (iii) an approach referred to as "reprogramming." I will present preliminary results for all of them and discuss advantages and disadvantages of the various approaches.

  本次报告将首先介绍音乐与音频分类任务的通用目标,之后将通过比较半监督(semi-supervised)、自监督表示学习(self-supervised representation)和重编程(reprogramming)等方法的优势与不足,结合初步的实验结果,重点论述数据的可获取性对于各项音频分类任务带来的挑战。

  Biography:

  Alexander Lerch is Associate Professor and Director of Graduate Studies at the School of Music, Georgia Institute of Technology,Co-Chair of ISMIR 2021. He received his ``Diplom-Ingenieur'' (EE) and his PhD (Audio Communications) from TU Berlin.

  Lerch teaches computers to listen to and comprehend music. His research positions him at the intersection of signal processing, machine learning, and music and creates artificially intelligent software for music generation, production, and analysis.

  Lerch authored more than 50 peer-reviewed journal and conference papers, as well as the text book "An Introduction to Audio Content Analysis" (IEEE/Wiley 2012).

  Before he joined Georgia Tech, Lerch was Co-Founder and Head of Research at his company zplane.development, an industry leader in music technology licensing. The technologies he worked on at zplane include algorithms such as time-stretching and automatic key detection. zplane technologies are nowadays used by millions of musicians and producers world-wide.

  Alexander Lerch 现任佐治亚理工学院音乐学院副教授兼研究生院主任,ISMIR大会主席。他于柏林工业大学获得了电子工程硕士和音频通信博士学位。Lerch教授的研究集中在信号处理、机器学习和音乐的交叉结合方面,训练计算机聆听并理解音乐,并开发了一系列可用于音乐生成、制作和分析的人工智能软件。Lerch教授撰写了 50 余篇经过同行评审的期刊和会议论文,以及教材《音频内容分析导论》(IEEE/Wiley 2012)。Lerch教授曾是音乐技术领域的领头企业Z-plane公司的联合创始人兼研发主管,在任期间参与研发了音频时间伸缩与自动音高识别算法等技术成果。Z-plane在全世界范围内拥有包含音乐家和制作人在内的百万级别的用户量。

  供稿:音乐人工智能与音乐信息科技系

  文:张渊、周子雅

 
 
  • 相关附件:
  • 相关链接:
XML 地图 | Sitemap 地图