Italian, adopted by France not long ago, I am a constant learner, dedicated to computer science and discovery—whether uncovering solutions or gaining insights.
The National Audiovisual Institute (INA) serves as the repository for all French audiovisual archives, maintaining a continuous archiving effort for over 180 radio and television services around the clock since 1995. The metadata generated from this vast collection equates to hundreds of billions of documents, encompassing images, audio and video fragments, and text excerpts.
Given the diverse nature of the content, the data model is inspired by the conceptual frameworks of cultural heritage, structured as an extensive graph with intricate relationships between generic entities.
Building a global search engine for this unique use case presents a dual challenge: ensuring rapid indexing and implementing sophisticated full-text search capabilities with high performance. The presentation will explore the crucial decisions made and the technical infrastructure developed to meet these challenges effectively, starting from data injection and indexing strategies, to the implementation of advanced search algorithms and scalable storage solutions.
Searching for speaker images...