AI for Video Games

I recently ran a course on Artificial Intelligence for Video Games. The slides can be found here:

http://fanafzargames.com/ahf/igdiai/

The course was an introduction to the applications of AI in video games and diverse topics were covered.

Follow the light

Akumana and Vahumana fight a lot in my daily life experiences. In Persian mythology, Akumana is a Deev which represents bad and evil thoughts and Vahumana (Bahman) is an Ahura representing good and positive thoughts.

I tend to believe that Vahumana is usually the winner but Akumana tries hard. My thoughts about the classes that I teach isn’t void of the battle between these two, especially whenever I feel that the majority in the class is with ignorant and un-motivated students. I either want to be teaching in a class with all my heart or rather not bet there, its hard for me to just put up a show. The whispers of Akumana have been mostly in the theme of despair, “this class is going nowhere… no one is learning anything… you can’t be changing anyone… its all a waste of time… nothing can be done in this system and the foundation of the education system is all corrupt…” and then there is Vahumana on the other side saying “ be patient…every little step counts… even if you can have a positive influence on one student, that is still great… this whole thing really has meanings and it will get better as time comes… be patient”.

There was a huge battle very recently between the two and Akumana was dominating for a whole week, while Azidahak was watching from the distance, before I decided about the new course and right when I was about to cancel all the classes for the upcoming semester, while I was sitting in our office alone on a beautiful sunny spring day, there came Vahumana, the strongest Vahumana I’d seen for years, gave me his sword followed by a few words in my ear and there I was, slashing the dark Akumana with the sword and prepared for the new experience, with a blue necklace of hope and a white robe of patience and the shining sword of optimism.

We go Vahumana, we go one more time, I’m no your side …

The concept drawing on the top represents Akumana and the one on the bottom is an illustration of Vahumana as they appear in the game of Soshiant.

Down Times and Wonders ...

One thing which brings me down a lot of times is when I see the wrong path people are going, try to avoid them, and not being able to do anything positive eventually. This happens to me a lot while I witness the progress of the students I teach in classes. Now I don't necessarily mean that I know whats best for people and their optimum paths in life, but in the small domain of education, when people have goals for themselves and they believe in those goals, then helping them move in that direction is a priority for me and not seeing this happen is quite sad. I usually don't feel sorrow for things that are not in my control and happen outside my domain but in this specific issue, I usually think that there are still lots of other things that I can do to have better effects on these people. I often feel that being a teacher is one of the most critical jobs in this world. Thinking that a few moves and reactions might have big consequences on important issues for people is just frightening sometimes. These effects can originate from any person anywhere but the academic domain magnifies it and provides an arena for the possibilities of affecting people.
I'm experiencing some down times currently.
It wasn't long ago that I was a student myself and thought about what teachers do and how they should be and how I can be ideal if I become a teacher one day. How you see all the problems and wrong doings of the teachers and you think you know all the solutions but once you cross the line and start performing from the other side, everyday you see that a lot of these issues are not trivial at all and the more you witness the years go by, the more you feel that maybe what all those teachers did was at last the best thing that can be done and thats how things should be and the reality is far from the ideal visions.
Actually I haven't quite given up in this regard and the above belief has not quite conquered me yet and I'll try to experience and realize my ideal or semi ideal beliefs through academia while I can... there are still more things that can be tried ... although a lot of down times can be generated in this struggle too.

Most of the complexity arises from the fact that you have to deal different with different people and what you do to have positive effects on X has probably negative effects on Y. So it seems like in order to be most effective, you need to ride a spiral and find a dynamic solution to this complex problem. This fact needs solutions that are radically different from what happens in education right now.
One of the main questions here is, should you focus on being best for a specific group or being average for most? Some believe that a good teacher should be the best for all, not focusing on a specific group, I don't agree with this sub-consciously but I'm not sure if I'm right. Sometimes I feel that it is the elite few that need most care and they are the ones that can make things happen and by supporting this group, great things can happen in the world, the energy spent for the common group might not lead to anywhere. At other times I think that maybe this elite group can do good on their own and small effects on the common group can have good effects for the community, since they are the majority. Not sure about the correct answer yet. If there are any correct answers at all.
Another big question is the effect of discipline in education. Should it be used in a strict manner or should the students be rather free and follow a liberal experience in education? To what degree is discipline needed? When is it too much? How much will you lose by not following proper discipline or forcing too much of it? This parameter is one that I've been slanted towards linearly more and more as years go by, I used to think that zero discipline is ideal and the way to go but have believed in it more and more as time goes.

I might have some answers some day ...

Educationopia

Another boring meeting with a government client. I had a good time to do a small simulation research meanwhile.

A little study goes like this:

A very simplified education system is consisted of 3 entities in the education area directly and one related indirectly. The 3 direct ones are "Student (S)", "Teacher (T)" and "Administration (A)". The indirect entity is "Industry (I)".

State Space :
Motivated Student: S+ Un-motivated Students: S-
Motivated Teacher: T+ Un-motivated Teacher: T-
Motivated Administration: A+ Un-motivated Admin: A-
Motivated Companies: I+ Un-motivated Companies: I-

Motivated student is trying to learn more and more, a motivated teacher tries to teach better with a higher quality, motivated admins want to have happy students and happy teachers and ... we'll see what else is needed for these admins in a bit.
Motivated companies try to increase their efficiency and quality of their work everyday.

Lets see what kind of educational institutions can exist.

1	2	3	4	5	6	7	8
s-	s-	s-	s-	s+	s+	s+	s+
t-	t-	t+	t+	t-	t-	t+	t+
a-	a+	a-	a+	a-	a+	a-	a+

Case 1 : Bad students, bad teachers, bad administrators. Well the students don't want to learn, the teachers don't want to spend lots of energy, usually very little content is delivered, really easy exams, rather good grades for the students, the students are happy, the teacher is happy, the admin can be happy or maybe not happy, if it is an a-, then it will be happy. So such a center is stable.

A stable center is one where everybody is happy and hence no changes happen to it in the future.

Case 2: Bad students, bad teachers. The admin is a good one. A good admin will not be satisfied with the situation and will probably try to change the teachers or the student intake. Can lead to case 8 in this situation.

Case 3: The teacher cares, but the students will have bad results, the admin is not happy that the students are not happy and will probably say goodbye to the teacher. Non stable and will move towards Case 1.

Case 4: Bad students in a good environment. Not stable, the students will be replaced, will lead to case 8.

Case 5: Good students in a bad environment, not stable. They will either let go of this institute or turn into s- in order to be happy. Converges to Case 1.

Case 6: Only bad teachers. Admin will know this very quick and probably replace the teacher. Becomes Case 8.

Case 7: Good students and teachers, administration not really good. Doesn't support the teachers, doesn't care about special requirements, will probably lose the teachers and then converge to Case 1. Students and teachers can't change the admins.

Case 8: Educationopia. Wow. Highly stable.

Now the main question is, how can an administration be a good one or a bad one? Well the first aspect was that a good administration wants to have happy students and teachers. But if this is the only case, then it would have problems with case 2, a good admin should not be silent in case 2.

It seems like a very good criteria for a good admin can be how well they know the industry and the requirements. If you have un-motivated teachers and students and they all get good grades at the end, how do you know your institution has problems? Well maybe you can be in touch with the companies in the industry and be able to judge based on their requirements. In that case you can easily see that although everybody gets good grades in your school, they have problems in the real world. Here is how it goes after the education period.

s+ s- s+ s-
i+ i+ i- i-

Motivated students become motivated individuals and professionals and if hired by a high quality company, everyone will have a good time and flourish, everybody happy and stability.
Un-motivated individuals will be rejected by good companies, if in time all they can get is un-motivated individuals, then the company turns into an un-motivated company after some time (After all, what is a company other than its people?).
Good individuals will not stay in bad companies and bad individuals love an un-motivated company and everything is stable.
So assuming that initially the number of i+ and i- are the same, if in time more s- are added, then we'll have more i-'s and if the input of s+ is more, then the i- can convert into i+. So we need more than 50% s+ coming out of the educational institutions.

In the model above if the initial distribution for s and t are 50% each, then the only way that we can have more than 50% output for s+ is that if case 1 turns to case 2. If that happens then 1 becomes 2 and 2 can become 8 and stable after some time.

So dear administration, if you want your society to be successful, please don't be satisfied with the happy faces of your students and the frowns of your teachers, see what's going on out there and judge based on that.

Maybe some time has passed by and the only I's out there are i-'s. This can happen, in that case when a tries to evaluate its performance with i's, it is not going to get valuable results and hence will remain an a-. In this case our beloved a should go global and measure everything in a global scale, shouldn't be that hard these days with all the information flowing everywhere.

I had to talk at this point and wasn't able to delve into the subject anymore.

Ending

It all ended, was very useful and informative. However the flight back to dubai is canceled due to lots of fog in Dubai and we will have to get back with one day delay. Lots of time to blog now.

Anyway probably the most interesting discovery for today came from a presentation that we got in the Malta NCC center about how they prepare their lecturers for the courses, mostly related to the practices the lecturers can follow to improve their efficiency in the class and being able to transfer the information to the students. The topic was really really interesting. I will have to reflect on it as much as I can later on.

One of the ideas in the presentation was that 90 percent of the knowledge that we have is in our subconscious, when we try to teach something, we are usually trying to be conscious about what we are talking about and this isn't easy. This is why we have many good professionals who aren't necessarily good teachers. The more you become an expert in a field, the more it moves into your subconscious. (The samurai have a good statement about this fact related to the definition of a Meijin, a master samurai who will engage in a war in a subconscious manner) And also this is why a lot of sub conscious actions of the teacher are important in the whole process, like the way he talks, his posture or body language, maybe sometimes more than what he is trying to say.

Other interesting topics were learning models and how it is important for the lecturers to know themselves first, the type of learner they are and then be able to understand what kind of weak points they might have, and how they can improve it in order to provide a solution for all types of learners in the class.

I'd have to launch a big research campaign now that I have found proper systematic information about these areas. I might be back with more later. :)

Rain and Sun at the same time is very popular in Malta.