Lecture Notes: Artificial Intelligence (AI) - Introduction

Course Introduction and Housekeeping

• Lecturer struggles with the microphone.
• Agrees it will be an interesting course, notes trends in student names.
• Class policy: No laptops allowed.

Introduction to Artificial Intelligence (AI)

• Overview: What AI is and its relevance.
• Course turnover: 10% change in student roster expected.

What is AI?

• Broad definition: Involves thinking, perception, and action.
• Relation to Models: Important in model creation—differential equations, probabilities, simulations.
• MIT approach: Model-based learning, even in humanities.
• Outcome for students: Will understand their own thinking better, build models of thinking, perception, and action.
• Representation: Essential for model making.
• AI Models: Representations that facilitate understanding, algorithms for thinking, perception, and action.
• Gyroscope example: Illustrates importance of representation in understanding physical phenomena.
• Representation of problems: Farmer, fox, goose, and grain problem as an example.

AI Problem-Solving and Methods

• Generate and Test method: Example with leaf identification.
• Importance of naming concepts: Rumpelstiltskin Principle.
• Difference between trivial and simple ideas: Simple can be powerful, don’t dismiss simple ideas as trivial.

The Role of Language and Visualization

• Visuomotor system: Involved in problem-solving (e.g., number of African countries equator crosses).
• Perception and action loops: Integrated to create intelligence; relevant AI concepts.

History of AI

• Lady Lovelace: First programmer, early ideas dismissing AI creativity.
• Alan Turing (1950): Turing test, foundational work in AI.
• Minsky (1960): Key paper titled “Steps Toward Artificial Intelligence.”
• Early AI programs: Jim Slagle’s symbolic integration, Eliza (conversational AI), other symbolic reasoning programs.
• Rule-based expert systems: Like Mycin (medical diagnosis), impactful in various fields.
• Deep Blue era: Computational power over intelligence.

Current AI Trends

• Integrated perception, action, and reasoning: Evolving understanding of intelligence.
• Example: Program using visual memory to answer questions.
• Broad influences: Cognitive psychology, linguistics, philosophy.

Future of AI

• Human intelligence and evolution: Initial human advancements in thinking and language.
• Language’s role: Enables storytelling, imagination, and perception; central to future AI development.

Course Logistics

• Lectures: Introduce material and big ideas; experience-oriented.
• Recitations: Expand material, facilitate discussion.
• Mega recitations: Focus on past quiz problems, problem-solving techniques.
• Tutorials: Assist with homework.
• Attendance: Correlation between lecture attendance and grades.

Grading System

• Conversion: From 0-100 scale to MIT’s 5-point scale based on understanding.
• Reassessment: Max of quiz score and corresponding final exam part.
• Final exam: Comprehensive yet time-pressurized.

Communication and Organization

• Tutorial sign-up: Immediate organization of student schedules.
• Mega recitation this week: Likely covering Python review, especially for those observing religious holidays.

Fill out the forms for tutorial scheduling before leaving. Further communication and updates will be provided online.