Teaching Notes - Midterm presentations in 'Language and Computers'

This post is a continuation from my previous posts on teaching a 100-level undergraduate course called Language and Computers. As mentioned earlier, it is a very diverse class, and I use this textbook: Language and Computers by Marcus Dickinson, Chris Brew and Detmar Meurers.

I cannot believe it is already half-done, but we had our mid-term presentations last week. I thought about how to do these mid-terms. We were half-way through, they had 3 assignments for the first 3 topics, and the fourth one was due after mid-terms. So everything discussed so far has been covered - why bother about a mid-term? Doing one-one oral examination (class of 25) is not a bad idea, but I wanted them to work in groups and expand on what we discussed so far. So, I came up with this list of 10-15 topics and divided the class into 6 teams.

3 teams chose DuoLingo, 2 teams chose regular expressions, and 1 team chose speech recognition which was not in my list. This week (2 classes of 50 min each - third class was revision class) has been the most enriching so far for me, for all the teams went beyond what was in the textbook (totally beyond - not even 10% overlap!) and presented about stuff I did not know.

I wondered if there will be repetitions since 3 teams are presenting on DuoLingo and it is a 100 level course. But we got to know how DuoLingo shows exercises for - English, Chinese, German, Greek, Hebrew, Italian, Japanese, Russian, Spanish and Ukranian :o. I learned how similar or different the language exercises and their sequencing are across languages (e.g., verb conjugations exercises seen in Spanish in early levels, but Cyrillic script based stuff in Russian), what happens when we go to higher levels (quite surprised to know DuoLingo still shows only short response exercises) etc and how this compares with TAGARELA described in the textbook. This was very englightening for me, and I suppose for them too, because the textbook chapter on Language Tutoring Systems did not have this richness.

Regular expressions - again, considering I had a feeling of guilt that I did not do the lab-exercise part well in the class, I wondered what to expect. But one group focused on using regular expressions for searching files, code etc (live demo) and the other group showed us a MS Word Macro they wrote with custom regular expressions! Both the groups had CS students or students with programming background and I am glad they made use of that to enhance our knowledge :)

The Speech recognition group presented a student of the Dictation tool on MacOS, talking about how accurate it is with different accents, and vocabulary. What I found really cool is the use of spectrograms to illustrate how the same word in different accents looks in terms of speech signal.

I ended the mid-term week with an review of topics so far, a preview of what to expect next, and took some feedback on things so far. Around this time, I wrote a short-script to randomly generate one name each time I ask a question. It worked hilariously on the first day - with some students who did not pay attention having to answer questions. I was a bit skeptical about continuing it, but decided to just do it nevertheless eventually.

Our next topic is text classification and we are in the middle of it right now.

Written on October 21, 2017