Sessions Include Language Metadata, Machine Learning, DNA-Based Storage
The 2019 Digital Asset Symposium (DAS), presented by the Association of Moving Image Archivists (AMIA), has announced a full day of speakers and presentations. DAS returns to the Museum of Modern Art (MoMA) in Manhattan on June 5.
DAS provides a unique opportunity to compare a wide variety of approaches from an array of institutions, organizations and individuals dealing with the same media challenges. Approaching the challenging topics from a case study approach as well as in presentations, DAS tackles the cutting-edge technologies as well as the real-world issues facing the community.
Nick Gold, DAS chair, “It’s about time. There are the limitations of time that we are given to achieve our goals; racing against the clock to make sure our assets are preserved with a limited amount of time available to us to fulfill our mission. And don’t forget the time we spend to make sure we are up to speed on the latest lingo and jargon!”
“Moving imagery is a celebration of the temporal domain, and we as a community are charged with overseeing this temporal resource over the course of longer periods of time. Compressing the amount of time to store information in new mediums and making sure we are working in the right ways to preserve history that can stand the test of time is the goal of DAS. Come spend the day with us to continue to learn and grow. It won’t be a waste of your time,” added Program Co-chair Jamie DiVenere.
Each year, distinguished speakers drawn from a variety of roles and organizations present topics of community interest. The 2019 DAS lineup includes:
DNA-Based Storage and Computation of Digital Data
Hyunjun Park, co-founder and CEO of Catalog DNA
Billions of years of evolution have selected for characteristics in DNA that make it an ideal medium for data storage and parallel computation. For example, DNA has one million times the data density of SSDs with a shelf-life of thousands of years. We as human beings, are examples of massively parallel DNA-based computers that make trillions of truly simultaneous calculations each moment. Harnessing the power of DNA for digital data storage and computation has become a topic of great interest, therefore, in the current age of explosive data creation and processing. CATALOG is an early stage startup founded by two MIT scientists, building the first cost effective DNA-based platform for these applications and launching pilot projects on it in 2019.
Tyranny in Triples
Lizzy Jongma, Senior ICT Project Leader Dutch Network for War Collections WWII
Tom De Smet, Deputy Director Netherlands Institute for Sound and Vision and Director of the board of the Dutch Network for War Collections
What can we learn from the blackest pages of European History?
In the Netherlands the Dutch Network for War Collections brings together digitized source material about WW2 and connects resources using Semantic Technologies. The Netherlands has a long tradition of Open Data Projects. Within the Network museums, libraries and Archives are now entering a new stage where they are beginning to combine collections and create user friendly, human-understandable web portals and digital channels for these materials.
Tryanny in Triples explores the use of new technologies to explain old data and “ancient” histories, from ethical issues (why publish Nazi Propaganda on Wikimedia Commons) and Big Data Analysis to the linking of names and recreation of lives (instead of graphs). Using thesauri to connect collections and why Sparql Endpoints are not the ideal approach.
Miles and Miles of Texas: Preserving Millions of Feet of Tape from Austin City Limits
Amanda Moore, Technical Project Director, KLRU TV, Austin
James Cole, Video Specialist – Content Manager, KLRU-TV Austin
Austin City Limits (ACL) is the longest running music show in television history and proud recipient of the National Medal of Arts and a Peabody Award. Performances at ACL have featured hundreds of legendary artists and bands, including Willie Nelson, Loretta Lynn, Ray Charles, Roy Orbison, Alison Krauss, Neil Young, LCD Soundsystem, St. Vincent, and Kendrick Lamar. ACL is produced by (and aired on) KLRU, Austin, Texas’s PBS station.
Throughout its 45 seasons, the producers of the show have recorded thousands of miles of video and audio tape, resulting in thousands of hours of tapes to digitize, catalog and make accessible as well as hundreds of terabytes of data to store and properly back up and manage. The ACL Archive team just completed the migration of all of the show’s video assets, gathered the majority of associated data and is on to phase II, digitizing multitrack and broadcast master audio tapes and will discuss how all of this is done with a core team of 2, on a public TV station budget, while still in production.
Fishing in your Data Lake – You Never Know What You’re Going to Catch
Dave Ginsberg, Visionary Technologist, Amazon Studios
Callum Hughes, Head of Post Production Technology, Amazon Studios
Amazon Studios generates petabytes of information yearly around all of the original show and feature content created. These content tributaries flow into a massive data lake that was created to categorize and manage all of these materials. But how do you manage millions of unstructured assets without standardized file naming or a directory structure? This is the opportunity that the Amazon Studios Post Production Technology team had in relation to finding a solution for their asset management at scale.
This session will examine how building backwards from what the customers of the data needed helped to create a flexible solution that utilizes machine learning to categorize materials. The novel solution is fully Cloud architected and uses many off the shelf AWS technologies to allow the Studio to leverage materials created across multiple content creation needs.
John Footen, Managing Director, Media & Entertainment, Deloitte
The technical side of our business has a seemingly endless supply of acronyms and buzzwords. Enough to play buzzword bingo? Of course! This is your chance to play buzzword bingo for real! Compete with your colleagues for the synergistic benefits of getting the most TLA’s without the benefit of machine learning or AI. Don’t miss this first of a kind game show/technical presentation!
Demystifying Language Metadata
Yonah Levenson, Manager of Metadata and Taxonomy, HBO
Broadcast and media companies are dealing with multiple language standards, typically using both proprietary and non-proprietary terms. Global distribution requirements are exposing the need to consistently present languages for audio, text on the screen, packaging/display, and online user interfaces. This presentation explores how the IETF BCP 47 languages standard provides a way for HBO and others in the media and entertainment industry to ‘herd the cats’ of existing system interoperability while scoping for a future state that allows better support of global efforts.
Machine Learning is Awesome, but it’s Still (and Always Will Be) About User Experience
Rodrigo Coelho, InSight Product Manager, Iron Mountain InSight
It is about adding structure to unstructured data. That’s a big deal when in working with Digital Archives. The good news is that now Machine Learning and its sophisticated algorithms can label objects, recognize faces, transcript what is being said, detect landmarks, and much more. Those generate a significant amount of metadata to be worked with. But metadata itself is not enough.
The goal is always to improve the experience of how assets are managed. It’s all about how to navigate and curate them, as easy and fast as possible. Metadata helps. But we also need to find an effective and visually way of handling large amounts of videos, images, documents, through a powerful User Interface that makes our work easier, less painful. Better user experience for us means better content for our customers.
Machine Learning is Awesome, will focus on work with the San Francisco Giants, organizing and curating their Digital Archive.
The Digital Asset Symposium is made possible by Preservation Partners Iron Mountain and Western Digital; Gold Sponsors ImagePro and LAC Group, Silver Sponsors DFT, Deluxe, and Prasad; and Bronze Sponsors Fotokem, Kodak, and NBCUniversal StudioPost, Fujifilm, AVP, and Chesapeake Systems.
To register for DAS or for more information, visit http://www.digitalassetsymposium.com.