X
X
X

X
Courses » AI: search methods for problem solving

AI: search methods for problem solving

ABOUT THE COURSE

For an autonomous agent to behave in an intelligent manner it must be able to solve problems. This means it should be able to arrive at decisions that transform a given situation into a desired or goal situation. The agent should be able to imagine the consequence of its decisions to be able to identify the ones that work. In this first course on AI we study a wide variety of search methods that agents can employ for problem solving.

In a follow up course – AI: Knowledge Representation and Reasoning – we will go into the details of how an agent can represent its world and reason with what it knows. These two courses should lay a strong foundation for artificial intelligence, which the student can build upon. A third short course – AI: Constraint Satisfaction Problems – presents a slightly different formalism for problem solving, one in which the search and reasoning processes mentioned above can operate together.



INTENDED AUDIENCE

This is a first course on Artificial Intelligence. While the intended audience is both UG and PG students studying Computer Science, in fact anyone comfortable with talking about algorithms should be able to do the course.


PRE-REQUISITES
Exposure to data structures and programming and an ability to discuss algorithms is the only pre-requisite.


INDUSTRY SUPPORT

Any industry that is involved in development of AI applications. This not only includes software companies (like Microsoft, Google, and Facebook) but also manufacturing companies like Ford and General Electric, and retail companies like Amazon and Flipkart.

7379 students have enrolled already!!

COURSE INSTRUCTOR



Deepak Khemani is Professor at Department of Computer Science and Engineering, IIT Madras. He completed his B.Tech. (1980) in Mechanical Engineering, and M.Tech. (1983) and PhD. (1989) in Computer Science from IIT Bombay, and has been with IIT Madras since then. In between he spent a year at Tata Research Development and Design Centre, Pune and another at the then youngest IIT at Mandi. He has had shorter stays at several Computing departments in Europe.


Prof Khemani’s long-term goals are to build articulate problem solving systems using AI that can interact with human beings. His research interests include Memory Based Reasoning, Knowledge Representation and Reasoning, Planning and Constraint Satisfaction, Qualitative Reasoning, and Natural Language Processing.


COURSE LAYOUT

Week  Topics

1 Introduction: Overview and Historical Perspective, Turing Test, Physical Symbol Systems and the scope of Symbolic AI,   Agents.

2 State Space Search: Depth First Search, Breadth First Search, DFID

3 Heuristic Search: Best First Search, Hill Climbing, Beam Search

4 Traveling Salesman Problem, Tabu Search, Simulated Annealing

5 Population Based Search: Genetic Algorithms, Ant Colony Optimization

6 Branch & Bound, Algorithm A*, Admissibility of A*

7 Monotone Condition, IDA*, RBFS, Pruning OPEN and CLOSED in A*

8 Problem Decomposition, Algorithm AO*, Game Playing

9 Game Playing: Algorithms Minimax, AlphaBeta, SSS*

10 Rule Based Expert Systems, Inference Engine, Rete Algorithm

11 Planning: Forward/Backward Search, Goal Stack Planning, Sussman’s Anomaly

12 Plan Space Planning, Algorithm Graphplan


The following topics are not part of evaluation for this course, and are included for the interested student. These topics will be covered in detail in two followup courses "AI: Knowledge Representation and Reasoning" and "AI: Constraint Satisfaction Problems".


A1 Constraint Satisfaction Problems, Algorithm AC-1, Knowledge Based Systems

A2 Propositional Logic, Resolution Refutation Method

A3 Reasoning in First Order Logic, Backward Chaining, Resolution Method




COURSE SYLLABUS

  • Overview and Historical Perspective, Turing Test, Physical Symbol Systems and the scope of Symbolic AI, Agents.
  • State Space Search, Heuristic Search, Solution Space Search, Stochastic Local Search, Population Based Methods.
  • Optimal Solutions, Algorithm A*, Admissibility of A*, Space saving variations of A*.
  • Problem Decomposition, Algorithm AO*, Rule Based Expert Systems, Rete Algorithm.
  • Game Playing: Algorithms Minimax, AlphaBeta, SSS*
  • Planning: Forward/Backward Search, Goal Stack Planning, Sussman’s Anomaly, Plan Space Planning, Algorithm Graphplan.

Text Book (Chapters 1-8): Deepak Khemani, A First Course in Artificial Intelligence, McGraw Hill (India), 2013


Certification exam:

  • The exam is optional for a fee. Exams will be on 22 October, 2017.
  • Time: Shift 1: 9am-12 noon; Shift 2: 2pm-5pm
  • Any one shift can be chosen to write the exam for a course.
  • Registration URL: Announcements will be made when the registration form is open for registrations.
  • The online registration form has to be filled and the certification exam fee needs to be paid. More details will be made available when the exam registration form is published.

CERTIFICATE

  • Final score will be calculated as : 25% assignment score + 75% final exam score
  • 25% assignment score is calculated as 25% of average of 12 weeks course: Best 8 out of 12 assignments.
  • E-Certificate will be given to those who register and write the exam and score greater than or equal to 40% final score. Certificate will have your name, photograph and the score in the final exam with the breakup. It will have the logos of  NPTEL and IIT MADRAS . It will be e-verifiable at  nptel.ac.in/noc.