Vector Databases and AI: Professor Mario Nascimento Discusses Challenges and Applications in an IIA Seminar

The Artificial Intelligence Institute (IIA) at LNCC invites you to the seminar "Approximate Nearest Neighbor Queries and Vector Databases (A Primer)", presented by Professor Mario Nascimento, a specialist in databases. The talk will take place next Wednesday, March 18, at 3:00 p.m., in a hybrid format: in person at Auditorium A at LNCC and live streamed on the institution’s YouTube channel.

But after all, what are vector databases? With the advancement of AI and language models such as ChatGPT, data increasingly needs to be represented mathematically through so-called embeddings. These representations are vectors in high-dimensional spaces, and managing them requires specialized systems: vector databases. They are essential for tasks such as personalized recommendations on streaming platforms, searching for similar images, and even enabling the functioning of virtual assistants.

During the lecture, Professor Nascimento will present the fundamentals of these systems, focusing on approximate nearest neighbor (ANN) queries. Unlike traditional searches, which look for exact matches (such as a CPF number), these queries identify similar items within large datasets. This makes it possible, for example, to find photos similar to a given image or articles related to a specific topic in just seconds.

The seminar will also address current challenges and solutions available on the market. The event is part of the “Artificial Intelligence Institute Seminars” series, organized by IIA/LNCC, which aims to disseminate knowledge about AI and its applications in society.