e. g.

You must have come across this term many times and you all must be aware of this term which means "for example". Technically its an abbreviation of "exempli gratia" in Latin. In the field of teaching, examples play a very important role. Many difficult concepts can be cleared with good examples. I also tend to give lots of examples during my lectures and students enjoy them a lot. I try to avoid to give examples given in the book and I always like to give my own examples. Sometimes they come to mind spontaneously or sometimes I have to think a lot. Recently I gave an example while teaching a topic called "Learning" in AI. Its just demonstrates how one can involve students to understand concepts like these with real life examples.



I was teaching Decision Tree and I got bored to take same examples of tennis-play or reptile-mammal which you find on internet. So I decided to take an example based on students' "taste" in photography. I decided that I would have 10 students to contribute training data and 4 students for test data. I asked them simple questions - whether they like scenery, whether they like photo with people, whether they like color/bnw photos. I recorded their answers in spreadsheet which I kept it open in front of them. I got the following responses -

Now that everyone got an idea of what students like in photography, I decided to show some sample photographs and ask the students to rate them. While doing so, I asked test data students(s11 to s14) to go out of the class for some time. I displayed two photos(p1 and p2) to first five students(s1 to s5) and next two photos(p3 & p4) to next five students(s6 to s10) and asked them to rate on a scale between 1(minimum) to 10(maximum). During this process, students s11 to s14 were completely clueless what is happening in the class. The ratings of the students are as follows -

After getting these ratings, I asked all the students to do a prediction based on these sample ratings for the students who were outside the class. Students by doing some analysis(like average), predicted the ratings that the outside students would give to these photos. There were differences of opinions among the students hence I recorded all the answers. They predicted the following -

Now the excitement reached at its peak as students realized what would be my next move! It was time to call the four students inside the class to give their answers who were wondering what must be happening in the class(Thank God that they didn't run away). I called them inside the class, told them that they would be shown photos which they have to rate. I kept our predictions hidden from them. The rest of the class were on the edge of their 'benches"! The students from s11 to s14 rated the photos like these -

There were lots of "Yes"s and "Ohh"s after these students gave their answers. The students who gave their answers didn't realize why there was so much excitement in the class. Finally it was the time to compare what was predicted by the class and how many answers actually matched.

The students who just gave answers were surprised to see that their answers were predicted correctly by the class! One answer didn't match at all. So that is when I introduced the concepts of bias, under-fitting and over-fitting. I also told them that what we did is a regression tree in which we predict numbers.

That day in the morning I was completely clueless of what I would be doing in the class. But after the lecture, I felt happy and excited! Somewhere I had read -

“Tell me and I forget, teach me and I may remember, involve me and I learn."

Its so true. Hope I would be able to do it many times!