Review: CS6601 Artificial Intelligence

Objective

The objective of taking this course was to learn the fundamental concepts in the field of Artificial Intelligence.

Learning Outcome

I thoroughly enjoyed taking this course. The assignments on Search algorithms and Game Playing algorithms improved my skills in Data Structure and Algorithms. For example, the bi-directional and tri-directional search algorithms not only taught me about the similarities between queue based searches like BFS, Dijkstras, and A* search, but also provided me with an opportunity to familiarize myself with the heapq library (Priority Queue). Furthermore, exposure to Game Playing concepts like mini-max, alpha beta pruning, and iterative deepening improved my understanding of how algorithms balanced optimality with speed.

There were also several other interesting topics such as Bayes Nets, Constraint Satisfaction Problems, Guassian Mixture Models, Neural Nets, Decision Trees, and Hidden Markov Models.

Difficulty

This course required consistent effort and a greater number of hours spent a week. I estimate anywhere between 10 to 30 hours spent a week, depending on when there is a submission due. I like to get started as soon as I can, and leveraged on days off from work to complete each assignment. The midterms and finals were of medium to medium-hard difficulty. I managed to do fairly well for the midterms and in all of my assignments so I did not have to worry too much during the finals.

Support

Ed Discussion posts really helpful because students posted questions frequently. Answers and FYI information from the teaching assistants were also timely and helped throughout the course. I also appreciated the walkthroughs for each assignment. They were informative and helped to get me started.