Sunghee Yun (He/Him)

Co-founder & CTO @ Erudio Bio, Inc., Chief Business Development Officer @ WeStory.ai, Advisor & Evangelist @ CryptoLab, Inc., KFAS-Salzburg Global Leadership Initiative Fellow @ Salzburg Global Seminar, Adjunct Professor @ Sogang University, Advisory Professor @ DGIST, Global Advisory Board Member @ Innovative Future Brain-Inspired Intelligence System Semiconductor of Sogang University, Network Expert @ Gerson Lehrman Group

Abstract

This comprehensive lecture explores the fundamental concepts and architectural innovations that define modern artificial intelligence, tracing AI’s remarkable evolution from symbolic reasoning systems of the 1950s to today’s transformative large language models and generative AI systems. Drawing from unprecedented market evidence—including ChatGPT’s explosive adoption reaching 35 million users in just 5 months and NVIDIA’s 101% year-over-year growth with 71.2% gross margins—the presentation demonstrates how AI has entered an accelerated adoption phase that surpasses previous technology revolutions. Through detailed analysis of adoption curves, investment patterns exceeding $28.2 billion in cumulative funding, and performance metrics approaching human parity across diverse cognitive tasks, the lecture establishes that we are witnessing not merely technological hype, but a fundamental shift comparable to the Internet revolution with even greater transformative potential.

The technical core of the presentation provides an in-depth examination of the architectural breakthroughs that enabled this AI renaissance, particularly focusing on the Transformer architecture introduced in the seminal “Attention is All You Need” paper. Through mathematical exposition of scaled dot-product attention mechanisms, multi-head attention, and self-attention variants, the lecture demonstrates how these relatively simple yet powerful linear-mapping-based attention models revolutionized natural language processing by enabling parallelizable computation and handling arbitrarily long dependencies. The discussion encompasses the evolution from sequence-to-sequence RNN-based models to modern Transformer variants like BERT and GPT architectures, illustrating how these foundations support the massive parameter scales of contemporary LLMs—from GPT-3’s 175 billion parameters to the emerging models exceeding hundreds of billions of parameters.

Looking toward the future, the lecture addresses both the extraordinary opportunities and critical challenges facing AI development, including energy consumption concerns, hallucination issues, and the imperative for responsible AI governance. The presentation concludes with analysis of generative AI’s transformative impact across industries—from creative content generation and healthcare applications to scientific discovery and business automation—while emphasizing that we are witnessing only the “tip of the iceberg” of AI’s potential. Through this comprehensive technical and strategic overview, attendees gain both the mathematical foundations necessary to understand modern AI architectures and the broader perspective needed to navigate AI’s unprecedented societal and economic implications.