Let's Get Down to Business
This week, I felt I took a step into what it actually means to do research work as a computer scientist, specifically in the field of Human-Computer Interaction. First of all, I learned how important it is to be a part of a team. During our group sketching sessions, I was able to have a little time to let my brain focus on the broader aspects of what I’m trying to accomplish this summer, rather than just being totally involved in the tiny details of code implementation.
Because of this new mindset, I was able to communicate with my fellow researchers and see that many of our projects overlapped and were co-dependent (but in a good way). For instance, everyone working on the Community aspect of Looking Glass, myself included, had to come together and make sure we were all thinking along the same lines when sketching out our designs. We talked through our group decisions and came out with stronger ideas than we had going in.
As part of my project, I want to be able to automatically tag actions in a world, but I don’t necessarily want to just pull method names because they could be non-descriptive. I knew Michael is working on pulling specific actions out of user’s worlds, so now we can work together to figure out how to make that happen.
I have been doing quite a bit of reading this week as well. I’m trying to learn as much as possible about how our target age range (~11-14) uses search engines on the internet, and how they relate to computers in general. I am also learning about Natural Language Processing, to the point where I found a book on the subject and am working my way through it. I would like to know exactly how computers can use algorithms that mimic the human brain to understand what a conversation partner (or user with a search query) is saying or looking for. When thinking about all of the different biological and mental processes that I go through when interacting with another person, I can’t even imagine how to break them down into algorithms simple enough to use in my code.
My hope is that all of my reading will pay off and I will be better able to plan out how to code a perceptive and useful search function with human-like (that is, relevant and specific) responses. This work has made me realize just how fascinating human-computer interaction really is. I realize I have only been a part of the DREU for about a month, but I am starting to think this could be the field for me. Studying information retrieval can be a little dry, but its connection to natural language processing is what interests me.
When working on the plan for an automated tagging system, for example, Jordana helped me realize that computers can’t function as well as humans in some ways yet. When creating a world, a user may have a specific feeling or intent that they want to convey in subtle ways that a computer could not pick up on. I would be able to auto-tag a turtle or a rabbit, but could I auto-tag a contemplative turtle or a spiteful rabbit? I don’t think so, at the moment.
For now, I need to plan out how exactly I will accomplish all of my goals by the end of summer. I need to map out a weekly schedule of detailed work plans so that I will be able to dive right into my tasks, rather than hesitating with a lack of direction. But I think I will need to be ready to revise those plans as I go along. I had two goals to accomplish by next Friday, and I already finished one on my first day of working on it. Sometimes I am surprised by how many tasks are hiding within one goal, and sometimes I am surprised by how efficiently I can work when I have the right tools. I am excited to move out of the planning phase and get started on finishing my work for the rest of summer.