Class Information

Lecture: T, TH 1:15 – 2:30pm SCI 128
Lab A: F 1:15 – 2:45pm SCI 256
Lab B: F 3:00 – 4:30pm SCI 256

Professor: Lisa Meeden
Office: SCI 243
Phone: 8565
Office hours: W 1:30 – 3:30pm, or you can stop by whenever my door is open.

Introduction

Artificial Intelligence (AI) can be defined as the branch of computer science that is concerned with the automation of intelligent behavior. Intelligent behavior encompasses a wide range of abilities, and as a result AI has become a very broad field that includes game playing, automated reasoning, natural language processing, modeling human performance, robotics, and machine learning. This course will explore a subset of these topics. In the first half of the semester we will focus on how to formulate problems in terms of search. We will create programs that use search to solve puzzles and to play games against an opponent. In the second half of the semester we will focus on machine learning. We will create programs using genetic algorithms, artificial neural networks, and reinforcement learning that automatically improve with experience. We will also discuss robotics and create programs to control simulated robots to navigate a series of increasingly more challenging mazes to find a goal. Throughout the course we will explore how AI relates to the interdisciplinary field of Cognitive Science.

Readings

There is no textbook for this course. We will be reading excerpts from several textbooks as well as articles and chapters from the primary literature. All readings are available as links off the schedule below. To prepare for class do the readings for the week prior to the first class meeting of the week.

Goals for the course

• Be able to respond to the common objections to the possibility of achieving AI, such as the objection that a computer can only do what we program it to do and therefore cannot create anything new.
• Understand that there are a variety of representations used to encode knowledge in AI problems. These representations can be classified along a continuum from symbolic to subsymbolic. Symbolic representations are atomic and discrete while subsymbolic representations are distributed and continuous.
• Understand that there are a variety of paradigms for approaching AI problems including search, machine learning, and embodiment.
• Be able to compare and contrast the strengths and weaknesses of various AI approaches.
• Given a particular AI problem, be able to determine an appropriate representation and approach for solving it.

---

Grading

Grades will be weighted as follows:

5%	Class Participation
45%	Labs
25%	Midterm Exam Oct. 6, in class
25%	Final Exam Dec. 6, in class

Academic Accommodations

If you believe that you need accommodations for a disability, please contact Leslie Hempling in the Office of Student Disability Services, located in Parrish 130, or e-mail lhempli1 to set up an appointment to discuss your needs and the process for requesting accommodations. Leslie Hempling is responsible for reviewing and approving disability-related accommodation requests and, as appropriate, she will issue students with documented disabilities an Accommodation Authorization Letter. Since accommodations may require early planning and are not retroactive, please contact her as soon as possible.

Lab Policy

Labs will be assigned on Fridays, during the scheduled lab time, and will be due the following Thursday by midnight. You will turn in your labs using the handin63 program.

Late labs will only be accepted if you contact me at least a day before the deadline with a legitimate reason for needing extra time (such as an illness or needing to leave campus).

Even if you do not fully complete an assignment, you should submit what you have done to receive partial credit.

---

Schedule

WEEK	DAY	ANNOUNCEMENTS	TOPIC & READING	LABS
1	Aug 30		Foundations of AI A proposal for the Dartmouth Summer Research Project on AI J. McCarthy, M. Minsky, N. Rochester, and C. Shannon Computing machinery and intelligence A. Turing Fifty years of AI: From symbols to embodiment and back L. Steels	None
	Sep 01
2	Sep 06		Search Solving problems by searching S. Russell and P. Norvig, sections 3.1-3.4	1: State space search
	Sep 08	Drop/Add ends (Sep 09)
3	Sep 13		Informed search Solving problems by searching S. Russell and P. Norvig, sections 3.5-3.7 (starting on page 92)	2: A* search
	Sep 15
4	Sep 20		Game playing Adversarial search S. Russell and P. Norvig A gamut of games J. Schaeffer	3: Minimax search
	Sep 22
5	Sep 27		Local and parallel search Beyond classical search S. Russell and P. Norvig, pages 120-126 Analagoy-making as a complex adaptive system M. Mitchell	Game tournament
	Sep 29
6	Oct 04		Review	None
	Oct 06	Midterm Exam
	Oct 11	Fall Break
	Oct 13
7	Oct 18		Machine learning Machine learning: Introduction T. Mitchell	None
	Oct 20
8	Oct 25		Evolutionary search Genetic algorithms: An overview M. Mitchell	4: Genetic algorithm
	Oct 27
9	Nov 01		Neural networks Artificial neural networks T. Mitchell	5: Classification
	Nov 03	Last day to declare CR/NC or W (Nov 04)
10	Nov 08		Reinforcement learning Reinforcement learning T. Mitchell Temporal difference learning and TD-Gammon G. Tesauro	6: Learning Tic-Tac-Toe
	Nov 10
11	Nov 15		Embodiment and Robotics Elephants don't play chess R. Brooks	7: Subsumption
	Nov 17
12	Nov 22		Natural intelligence vs Artificial intelligence What can AI get from neuroscience? S. Potter	None
	Nov 24	Thanksgiving Break
13	Nov 29		Successes of AI Building Watson: An overview of the DeepQA project D. Ferrucci et. al. DARPA Grand Challenge Review	Subsumption Challenge Results
	Dec 01
14	Dec 06	Final Exam

Links

• AI in the news
• Expo of Robotics Research