VAM! AI Reading Group - Paper: TurboQuant

​📄 Paper: TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate

​https://arxiv.org/abs/2504.19874

​TurboQuant is a vector quantization algorithm that compresses high-dimensional vectors to low bitwidths while provably preserving their geometric structure. It's data-oblivious, meaning it works in real-time without any calibration data, and it gets within ~2.7x of the theoretical best possible compression for any bitwidth.

​It matters because compressing vectors without losing what they mean is a foundational problem across modern AI - from running LLMs efficiently to scaling search and retrieval systems - and this paper is one of the first to solve it with formal guarantees.

​If you are brand new to AI, we will spend the first 10ishg minutes reviewing the most important aspects of the Transformer Architecture, to help give you a background understanding of the topic before we dive into this paper.

​Presenter:

​Pranav is an Adjunct Professor at UBC Engineering and the Program Director at TKS, where trains young people how to build with emerging technologies (AI, Brain Computer Interfaces, Genetics etc) to solve important problems.

​Recommended Action Item:

• ​Given the complexity of this paper, it is not expected that you understand all the technical details. However, it is helpful to spend sometime understanding the first principles of it using AI as a learning resource

​Want to present?

• ​Please reach out to Issam!

​To connect with the group, join the Discord: https://discord.gg/teJvEejs94

​Timeline:
🕠 6:00 PM – Arrival & Networking.

​🗣️ 6:10 PM ~ 7:30 – Paper Presentation & Discussions

​About the Founder of VAM

​Issam Laradji is a Research Scientist at ServiceNow and an Adjunct Professor at University of British Columbia. He holds a PhD in Computer Science and a PhD from the University of British Columbia, and his research interests include natural language processing, computer vision, and large-scale optimization.

​Looking forward to discussing the latest AI Papers!

​About Northeastern University

​Founded in 1898, Northeastern is a global research university and the recognized leader in experiential lifelong learning. With 13 campuses across the U.S., U.K., and Canada, over 320,000 alumni, and more than 3,800 partners worldwide, the global university network serves as a platform for scaling ideas, talent, and solutions. Located in the heart of downtown, the Vancouver campus offers a wide range of master’s-level tech degrees, including pathway programs that allow students with non-STEM backgrounds to pursue a graduate degree in tech.