You are cordially invited to the Analysis Seminar organized by the Department of Mathematics.<\/p>\n
Speaker: Berkay Anahtarc\u0131 (\u00d6zye\u011fin University)<\/p>\n
“Mathematical Foundations of DeepSeek: A Reinforcement Learning Perspective.”<\/p>\n
Abstract: This talk explores the mathematical foundations of DeepSeek R1, an RL-only model designed for complex reasoning. Unlike traditional supervised fine-tuning, DeepSeek R1 leverages Group Relative Policy Optimization (GRPO), a novel method that stabilizes Proximal Policy Optimization (PPO) without a critic. GRPO enhances chain-of-thought reasoning by structuring problem-solving into sequential steps. I will analyze its theoretical properties and implications for reasoning-driven reinforcement learning.<\/p>\n
Date: Monday, February 24, 2025
\nTime: 17:00-18:00, GMT+3
\nPlace: ZOOM<\/p>\n
This is an online seminar. To request the zoom link, please send a message to goncha@fen.bilkent.edu.tr<\/p>\n","protected":false},"excerpt":{"rendered":"
You are cordially invited to the Analysis Seminar organized by the Department of Mathematics. Speaker: Berkay Anahtarc\u0131 (\u00d6zye\u011fin University) “Mathematical Foundations of DeepSeek: A Reinforcement Learning Perspective.” Abstract: This talk explores the mathematical foundations of DeepSeek R1, an RL-only model designed for complex reasoning. Unlike traditional supervised fine-tuning, DeepSeek R1 [\u2026] <\/p>\n