FrontierMath to Infinity - Mathematical Problem Solving Benchmark and AI Trajectory
Harvard CMSA via YouTube
Python, Prompt Engineering, Data Science — Build the Skills Employers Want Now
Master AI and Machine Learning: From Neural Networks to Applications
Overview
Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
Explore the development and implications of FrontierMath, a cutting-edge mathematical problem-solving benchmark, in this conference talk by Elliott Glazer from Epoch AI at Harvard CMSA's Geometry of Machine Learning series. Delve into the design philosophy behind FrontierMath and discover key insights about AI's current trajectory in mathematical reasoning capabilities. Examine what constitutes a "perfectly efficient" mathematical intelligence and learn how high-ceiling mathematics capability metrics can illuminate the pathway toward achieving that theoretical ideal. Gain valuable perspectives on the intersection of artificial intelligence and advanced mathematical problem-solving, with forward-looking speculation about the future potential of AI systems in mathematical domains.
Syllabus
Elliott Glazer | FrontierMath to Infinity
Taught by
Harvard CMSA