Artificial Intelligence: Foundations and Algorithms (Intro Video)

Added: 2026-06-02

[00:00:00]Hello everyone. Uh, welcome to the course titled Artificial Intelligence Foundations and Algorithms. My name is Mausam.

[00:00:08]I'm a professor at Computer Science Department and jointly affiliated with the Aarti School of Artificial Intelligence at IIT Delhi.

[00:00:16]Let me tell you a little bit about my career. I did my PhD at University of Washington in Seattle in 2007. And it feels like old ages actually now. And I stayed there for 6 more years and I was a research assistant professor there.

[00:00:30]And then I moved back to India in 2013, joined IIT Delhi and have been at IIT Delhi ever since.

[00:00:35]Um, I have been teaching a lot of AI courses.

[00:00:38]And this course, which is sort of the introductory course in AI, uh, I taught several times uh, every year and at some point I felt that the course had got a certain finesse, a certain amount of, you know, well-roundedness uh, to actually record it for NPTEL and I recorded it and the course has been running for many years now and uh, has had a very, very positive feedback.

[00:01:02]Almost 40,000 to 50,000 students take it every year and I have received a lot of uh, you know, good wishes from various students who have benefited from it. So, I'm very happy to share this again with you.

[00:01:13]Uh, let me tell you a little bit about this course. It's an undergraduate course. It's a 12-week course.

[00:01:18]Um, now AI as a field is really huge. It's really vast. There are just so many different perspectives to look at AI.

[00:01:25]And so, when we teach an introductory course, we have to make some choices and uh, the choice I have made in this course is that I will cover introduction to a lot of material, but I may not go as much into the depth of any of them, right? So, I will tell you the first or the first two two, three algorithms that you will learn in that field in that subfield of AI. But, if you want to pursue that further, then you have to go and take an advanced course in that particular area. We actually talk about the philosophy of AI and we do not yet only cover the the the things that have been happening in the last 5-10 years, right? We all know about AI. We all know how much it has been changing and revolutionizing the world in the last uh few years.

[00:02:09]But as a professor, it is important for me that not only do we cover some of the more modern aspects of AI, but we give a balanced view to our students and also tell them the 70-year history that AI has.

[00:02:21]So from that perspective, we start out by talking about what is AI. What does it mean for the machine to be intelligent? And how as humans we would even know about it if the machines exhibited intelligent behavior, okay? So that's really the first week. And then we talk about the traditional algorithms in AI, the traditional views of AI, which is the search view and the knowledge representation view. This is the view This is based on uh logic representation language. This is the view that was the popular view for 50 years in the field of AI, and we spend 5-6 weeks talking about that. Then we change sites. We get more modern. So more modern would be what has happened in the last 20 years. And we first take up probabilistic view, and towards the end we take the neural view. Uh another important thing to note is that often AI course is coupled with a machine learning course.

[00:03:10]And uh as AI uh I take the the perspective that I should teach you everything, but at the same time I should not be teaching you things that you will anyway learn in a machine learning course uh just after AI, right? So therefore, I don't cover a lot of the machine learning that you will learn about in a standard machine learning course.

[00:03:30]Instead, I teach reinforcement learning, which is actually something should be covered in machine learning, but is generally not covered. So that way after AI, if you take the machine learning course, you get a really really great perspective of all of what is happening in the field of AI, and then you can get into the more advanced topics in uh sort of deep learning and natural language processing and so on and so forth.

[00:03:51]We follow a book. Our book is Artificial Intelligence: A Modern Approach. It is written by Stuart Russell from Berkeley and Peter Norvig at Google. I always think of it as a sort of like a Bible of AI.

[00:04:04]If you If you have it, and this book is sort of like really big by itself. So, it is impossible in a course in a 12-week course to cover all of it. We sort of cover about 50% of the book. So, you get a really good understanding of what's happening in the field.

[00:04:20]Now, one important aspect I thought for this course is not just to teach you technical content, but also give you the wider perspective. So, what is the wider perspective? You know, there are stories about how AI came about. There are competitions that happened. Somebody won. Somebody expected to win, but did not win. There There have been lots of anecdotes about, you know, how AI researchers sometimes anecdotes about, you know, AI ideas really help us in our life or, you know, in understanding primate behavior. So, we tell you a lot of these stories. And these stories, these details become like a fun part of the course, and it not only give you They not only give you the perspective about the AI algorithms, but give you a deeper understanding of how AI algorithms can be connected to the world.

[00:05:05]Last, but not the least, I do want to say that when this course ran in the in its previous offering, lots of students took it, but fewer students actually formally finished the course. And we did some introspection and some uh you know, thinking about it, and we realized that by calling it a completely introductory course, the students get the view that they don't need to have any prerequisites to really understand this course. That's not true.

[00:05:31]So, if you are doing it just for fun, I obviously invite you to do it, and you know, you will learn something from it.

[00:05:37]But if you want to take it for credit, if you want, you know, a certificate of completion, then it is my suggestion that you should have background in data structures, basics of algorithms, uh, discrete structures, and probability. If you have that background, then you will be able to do well in this course, but if you don't have this background, then, you know, you might, uh, find the course a little bit too rigorous. So, this is just my suggestion. You can do self-selection. You're obviously welcome in the course, and I really want you to succeed, and I really want you to pass, but based on previous experience, I thought I would share these prerequisites. I hope you will enjoy this course. I, uh, if you have any suggestions, uh, for me, if any feedback for me, any content you found, you know, confusing, please let me know, and I'm happy to re-record those parts of the lecture, and, you know, get additional content for you to, you know, make it easier. But, I hope you have fun, and welcome to the course. Thank you.