Peeking Behind the Curtains of Large Language Models
Join us for an insightful journey into the world of Large Language Models (LLMs) such as ChatGPT. This talk will discuss the algorithmic and mathematical foundations of training these models on vast text corpuses, the processes behind prompt responses, and their applications in domains such as healthcare and education. We’ll also explore the compute infrastructure in large-scale data centers, and discuss the challenges and future directions in enhancing these models and expanding their accessibility. Perfect for anyone curious about the cutting-edge technology behind LLMs and their transformative potential.
About the Speaker:
Gohar Irfan’s research is on improving the resource efficiency of cloud computing infrastructure. This involves working on the boundary of software and emerging hardware technologies with the goal of maximizing performance of modern applications (e.g., artificial intelligence) while minimizing the resource usage (e.g., power, carbon footprint). Prior to starting his PhD, Gohar worked at Microsoft Research where he worked on several projects to improve resource efficiency of Microsoft’s datacenters.
This session is free and open to all. Just visit the venue to attend it.
The Black Hole
Plot 5H, Street 100, G-11/3, Islamabad.
Click here for Google Maps Location