AMPPD Project Update
Loading...
Can’t use the file because of accessibility barriers? Contact us with the title of the item, permanent link, and specifics of your accommodation need.
Date
2020-10-07
Journal Title
Journal ISSN
Volume Title
Publisher
Indiana University Digital Collections Services
Permanent Link
Abstract
Academic libraries and archives are dealing with increasing numbers of digital audio and video (AV) files, acquired through both digitization of analog collections and acquisition of born-digital AV resources. While the emergence of low-cost storage options and maturity of streaming platforms has made it easier to store and deliver AV, these collections often lack metadata needed in order to make them discoverable and usable by researchers and other users. Since late 2018, the Indiana University Libraries have been working with partners at the University of Texas at Austin, New York Public Library, and digital consultant AVP to develop an open source software platform, known as AMP (Audiovisual Metadata Platform), that leverages automated machine learning-based tools together with human expertise to build workflows to create and augment metadata for AV resources to improve discovery, rights determination, and use. We will present an update on progress of the AMP project and its successes and challenges to date, including a demonstration of the AMP system and discussion of issues in system design, workflows, and the use of open source and commercial cloud-based machine learning tools. We will also discuss results to date of testing the AMP system using collections from the Cook Music Library and University Archives at IU and from the New York Public Library. This work is generously supported by a grant to IU from the Andrew W. Mellon Foundation.
Description
Keywords
Audiovisual materials, Metadata, Library resource metadata, Machine learning
Citation
Journal
DOI
Link(s) to data and video for this item
Click the link below to play this video
Rights
Type
Presentation