인공지능 lecture01

Artificial Intelligence Course Introduction

Overview

• Professor Sang‑ming Lee, PhD from KAIST, introduces the AI course and his background.
• Course is held offline; attendance, assignments, and exams comprise the grading.
• Focus: provide a broad understanding of AI, machine learning, and deep learning, preparing students for advanced electives.

Key Concepts

• Artificial Intelligence (AI) – broadest field; techniques that enable computers to mimic human intelligence.
• Machine Learning (ML) – subset of AI; algorithms that learn patterns from data without explicit programming.
• Deep Learning (DL) – subset of ML; uses deep neural networks (multiple layers of perceptrons) to learn complex representations.
• Four AI Approaches
  1. Human‑like thinking (cognitive modeling)
  2. Human‑like acting (behavioral outcomes)
  3. Rational thinking (logical inference)
  4. Rational acting (goal‑oriented agents)
• Learning Phases – training (model learns from labeled data) and inference (model predicts on new data).

Detailed Notes

• Instructor Profile
  • Graduated Yonsei University; PhD from KAIST.
  • Former researcher at Georgia Tech, University of Illinois, and Sungkyunkwan University.
  • Research interests: multimodal learning, self‑supervised learning, social AI.
  • Contact: office in Science Library Building, Moon 609C; email provided.
• Course Logistics
  • Offline format; no use of Building 611.
  • Attendance counts for 10% of grade; each missed session deducts 1%.
  • Assignments: 20%; two exams (midterm and final): 35%.
  • Syllabus is tentative; includes search algorithms, reinforcement learning, ML basics, deep learning techniques.
• AI Definitions & Structure
  • AI encompasses all methods to emulate human intelligence.
  • ML learns from data and corresponding labels; DL specializes in neural networks with many layers.
  • Historical context: neural nets date to 1940s‑60s; modern success due to big data, GPU power, and improved algorithms.
• Applications Covered
  • Strategic Games: AlphaGo, Atari, StarCraft – reinforcement learning examples.
  • Computer Vision: image classification, segmentation, object localization, action recognition, video generation.
  • Natural Language Processing (NLP): machine translation, summarization, sentiment analysis; unified models like GPT, Gemini.
  • Audio Processing: speech recognition, music analysis, generation, audio enhancement, synthesis.
• Course Goal
  • Equip students with a comprehensive AI overview to enable them to pursue specialized courses such as reinforcement learning, deep learning, machine learning, computer vision, or NLP.

Takeaway: This introductory lecture sets the foundation for understanding AI’s scope, key subfields, and real‑world applications, positioning students for deeper study in advanced AI disciplines.