DeLTA seminar by Amin Karbasi: The Revolution of Light Language Models: Acceleration, Application, Adversary

Press here to join the Zoom call. 

Speaker

Amin Karbasi, Chief Scientist at Robust Intelligence and a professor at Yale University

Title

The Revolution of Light Language Models: Acceleration, Application, Adversary

Abstract

In this talk, I'll discuss how our research is trying to push the boundaries of generative AI with light language models. These models offer new possibilities due to their efficiency, low memory/energy consumption and potential for faster learning. I will particularely focus on three key areas: 1- Large Context: We are building light language models that can handle massive amounts of context, accelerating their ability to process information. 2- Brain Language Model: With faster, more efficient models, we will explore the possibility of them learning the complex language patterns used by the human brain, potentially offering insights into how we think and communicate. 3- Adversarial Attacks: We will see how light models can be leveraged to exploit vulnerabilities in much larger language models, potentially forcing them to generate unintended outputs.

_____________________________

You can subscribe to the DeLTA Seminar mailing list by sending an empty email to delta-seminar-join@list.ku.dk.
Online calendar
DeLTA Lab page