/
lectern.tv
Play15s30s1x0:00/--:--
DoneWatched
Expand
lectern.tv
51:33/55:57
1:01
2:12
5:10
2:43
1:34
6:23
3:05
4:25
7:54
8:29
5:57
4:43
2:21
Artificial Intelligence
Harvard's CS50 intro to AI: from rubber ducks and minimax to neural networks, LLMs, and prompt engineering.
Lecture guide · Written by AI from the lecture. For the lecture’s own words, open the Transcript tab.
Introduces the foundational concepts of artificial intelligence, covering tools like image generation, deep fakes, and text-based large language models. David J. Malan explains how CS50 integrates AI as a virtual teaching assistant through custom system prompts, and demonstrates core algorithmic concepts including decision trees, the minimax algorithm, reinforcement learning, and neural networks.
This lecture assumes
- Game Theorytaught in Introduction to Individual Decision-Making at 0:27
- Vector spacetaught in On Notation: Principles and Impact in Computing at 9:40
Key Takeaways
- AI tools like text generators and image synthesis are advancing rapidly, making misinformation increasingly difficult to discern.
- Prompt engineering and system prompts can configure large language models to behave like supportive, Socratic teachers rather than answer-dispensing shortcuts.
- Games like tic-tac-toe can be solved deterministically using decision trees and the minimax algorithm to maximize or minimize scores.
- Reinforcement learning trains agents through trial and error by rewarding positive behaviors and penalizing negative ones.
Concepts Explained19
Rubber duck debugging—A method of debugging code by explaining it aloud to an inanimate object, such as a rubber duck, to help realize the logic errors in one's thoughts.
Deep fake—A video synthesized by software using artificial intelligence that mutates an actual image or video of a person to fake their actions or words.
Prompt Engineering—The English instructions given to AI to shape its behavior, personality, and constraints as a teacher or duck.
System Prompt—A representative snippet of English sent to an AI service to teach it how to behave educationally and constructively.
Examples, Proofs & Derivations
Works & Sources Cited3
Paper
Robot Motor Skill Coordination with EM-based Reinforcement Learning—Petar Kormushev, Sylvain Calinon, and Darwin G. Caldwell, Italian Institute of Technology
Paper
Attention Is All You Need—Ashish Vaswani et al., Google Brain / Google Research / University of Toronto, 2017, introducing the Transformer architecture.
Book
Homework Machine—Poem by Shel Silverstein published in 1981.
Part of course·Lecture 8 of 18
CS50: Introduction to Computer Science
About the lecturer
Gordon McKay Professor of the Practice of Computer Science at Harvard University
Gordon McKay Professor of the Practice of Computer Science at Harvard University and lead instructor of CS50.
2 courses · 28 lectures in archiveView profile
Discussion
Share a seminar note, question, or timestamped reference (e.g. 14:20)...Sign in