Detailansicht
Deep Learning Techniques for Music Generation
eBook - Computational Synthesis and Creative Systems
ISBN/EAN: 9783319701639
Umbreit-Nr.: 8214298
Sprache:
Englisch
Umfang: 0 S., 12.16 MB
Format in cm:
Einband:
Keine Angabe
Erschienen am 08.11.2019
Auflage: 1/2019
E-Book
Format: PDF
DRM: Digitales Wasserzeichen
- Zusatztext
- <p>This book&nbsp;is a survey and analysis of how deep learning can be used to generate musical&nbsp;content. The authors offer a comprehensive presentation of the foundations of&nbsp;deep learning&nbsp;techniques for music generation. They also develop a conceptual&nbsp;framework used to classify and analyze various types of architecture, encoding&nbsp;models, generation strategies, and ways to&nbsp;control the generation. The five dimensions&nbsp;of this framework are: objective (the kind of musical content to be generated, e.g.,&nbsp;melody, accompaniment); representation (the musical&nbsp;elements to be considered and&nbsp;how to encode them, e.g., chord, silence, piano roll, one-hot encoding);&nbsp;architecture (the structure organizing neurons, their connexions, and the flow&nbsp;of their&nbsp;activations, e.g., feedforward, recurrent, variational autoencoder);&nbsp;challenge (the desired properties and issues, e.g., variability,&nbsp;incrementality, adaptability); and strategy (the way to model&nbsp;and control the&nbsp;process of generation, e.g., single-step feedforward, iterative feedforward,&nbsp;decoder feedforward, sampling). To illustrate the possible design decisions and&nbsp;to allow&nbsp;comparison and correlation analysis they analyze and classify more&nbsp;than 40 systems, and they discuss important open challenges such as interactivity,&nbsp;originality, and structure.<br><br> The authors&nbsp;have extensive knowledge and experience in all related research, technical,&nbsp;performance, and business aspects. The book is suitable for students,&nbsp;practitioners, and&nbsp;researchers&nbsp;in the artificial intelligence, machine learning, and music creation domains.&nbsp;The reader does not require any prior knowledge about artificial neural&nbsp;networks, deep learning, or&nbsp;computer music. The text is fully supported with a&nbsp;comprehensive table of acronyms, bibliography, glossary, and index, and&nbsp;supplementary material is available from the authors' website.</p><p></p>
- Kurztext
- This book&nbsp;is a survey and analysis of how deep learning can be used to generate musical&nbsp;content. The authors offer a comprehensive presentation of the foundations of&nbsp;deep learning&nbsp;techniques for music generation. They also develop a conceptual&nbsp;framework used to classify and analyze various types of architecture, encoding&nbsp;models, generation strategies, and ways to&nbsp;control the generation. The five dimensions&nbsp;of this framework are: objective (the kind of musical content to be generated, e.g.,&nbsp;melody, accompaniment); representation (the musical&nbsp;elements to be considered and&nbsp;how to encode them, e.g., chord, silence, piano roll, one-hot encoding);&nbsp;architecture (the structure organizing neurons, their connexions, and the flow&nbsp;of their&nbsp;activations, e.g., feedforward, recurrent, variational autoencoder);&nbsp;challenge (the desired properties and issues, e.g., variability,&nbsp;incrementality, adaptability); and strategy (the way to model&nbsp;and control the&nbsp;process of generation, e.g., single-step feedforward, iterative feedforward,&nbsp;decoder feedforward, sampling). To illustrate the possible design decisions and&nbsp;to allow&nbsp;comparison and correlation analysis they analyze and classify more&nbsp;than 40 systems, and they discuss important open challenges such as interactivity,&nbsp;originality, and structure. The authors&nbsp;have extensive knowledge and experience in all related research, technical,&nbsp;performance, and business aspects. The book is suitable for students,&nbsp;practitioners, and&nbsp;researchers&nbsp;in the artificial intelligence, machine learning, and music creation domains.&nbsp;The reader does not require any prior knowledge about artificial neural&nbsp;networks, deep learning, or&nbsp;computer music. The text is fully supported with a&nbsp;comprehensive table of acronyms, bibliography, glossary, and index, and&nbsp;supplementary material is available from the authors' website.