Machine Learning + Libraries Summit: Event Summary now live!

On Friday, September 20, 2019, the Library of Congress hosted the Machine Learning + Libraries Summit. This one-day conference convened 75 cultural heritage professionals (roughly 50 from outside the Library of Congress and 25 staff from within) to discuss the on-the-ground applications of machine learning technologies in libraries, museums, and universities. Hosting this conference was part of a larger effort to learn about machine learning and the role it could play in helping the Library of Congress reach its strategic goals, such as enhancing discoverability of the Library’s collections, building connections between users and the Library’s digital holdings, and leveraging technology to serve creative communities and the general public.

The Machine Learning + Libraries Summit Event Summary is now available as a downloadable report on labs.loc.gov. This document includes more detailed information about the conference proceedings. It broadly summarizes recurring themes of discussion and compiles the outputs of the small group activities. We hope it serves as a point of entry into broader conversations around the challenges, opportunities, and actionable items concerning machine learning in cultural heritage.

Read the full report here. The executive summary can be found below. Please share your feedback in the comments.

•••

The event was divided into three themes: 1) ongoing projects, 2) opportunities and challenges regarding partnerships & vendors, and 3) future applications. Each thematic strand included lightning talks, a small group activity, and a whole group discussion.

The goals of the conference were to:

• survey the range of ongoing projects in the broader cultural heritage landscape;
• surface major possibilities and barriers for applying machine learning in a library setting;
• demonstrate the possibilities of machine learning for use at the Library of Congress to internal audiences.

Threads emerging from whole group discussion at the conference include:

• ethics, transparency, and communication;
• access to resources;
• attracting interest in GLAM (galleries, libraries, archives, museums) datasets;
• building machine learning literacy;
• expanding machine learning user communities;
• operationalization;
• connecting machine learning and crowdsourcing;
• metrics for evaluation of vendors and projects; and
• copyright and implications for the use of content.

The small group exercises were organized around the following themes:

• Defining success in (machine learning) projects: The materials gathered from this exercise suggested that, at first blush, results and outcomes were identified as keys to “success;” however, further discussion surfaced the ways subject expertise is essential to practically implementing machine learning.
• Takeaways for collaboration on (machine learning) projects: Observations loosely fell under the broad themes of: 1) project management, 2) expectations management, 3) data, 4) resources, and 5) team composition.
• Milestones for machine learning projects in the next 6 months, 1-2 years, and 3-5 years: The most frequent asks were around developing educational programming for MLIS students and building technical literacies; the need for funding; and desired documentation to be created about best practices, use cases, and ethical considerations.