The Leaderboard Illusion | Beyza Ermiş, Cohere
Join the Bilkent AI Society's 10th Seminar featuring Beyza Ermiş, Research Scientist at Cohere, for a crucial discussion on AI benchmarking reliability and insights from her research career. Register now for free!
Beyza brings valuable experience from the forefront of AI research, currently at Cohere and previously spending over five years at Amazon Web Services (AWS). A Bilkent University alumna (BS, Computer Engineering), she holds a PhD from Boğaziçi University and focuses on advancing large language models and their trustworthy evaluation.
This session will dive into the findings of the impactful paper, "The Leaderboard Illusion," co-authored by Beyza. We'll explore how AI benchmarks can be distorted, echoing Goodhart's Law: "When a measure becomes a target, it ceases to be a good measure." Beyond the paper, Beyza will share valuable insights and learnings from her journey in leading industrial AI research labs.
Who Should Attend:
🤖 Enthusiasts of AI and machine learning
🧐 Anyone interested in AI evaluation, benchmarking, and research integrity
🎓 Students, researchers, and professionals in computer science and engineering
🚀 Anyone curious about the state of AI progress and cutting-edge research
📅 Date: May 7, 2025 🕗 Time: 17:30 (TRT / GMT+3) 📍 Location: Online (Register to get the meeting link)
🌟 Register now for free! https://register.bilkentai.com