Monday, September 26, 2016: 11:24am – 6:18pm (6.9 hours):
Professor had a meeting around noon, so our lab meeting (and most likely Kaushik and I’s presentation) will be pushed to Wednesday.
I spent hackathon implementing mode extraction from a .sql file. It’s a rocky start, but the gist is that it takes the “CREATE TABLE” sections of the file, then splits it into primary, secondary, and foreign keys. On paper, this seemed like it would be the solution, but translating this into first-order logic is still unsolved. I poked around on the internet for a little while and have suspicions that nltk may have something I need, but I’ll need to do a lot more digging.
Mayukh and I spent some time talking through the problem. We walked through an example and I told him a bit about the problem I was having. How to solve this problem is completely based on what assumptions you can make: I learned last week that the arity was easy to solve if you already had a facts file, and finding keys is easy if you have a .sql file with CREATE TABLES. It’s difficult to go from pure data (in the form of a txt or csv file) into predicates, facts, and finally: knowledge.
Tuesday, September 27, 2016:
Professor had a really interesting lecture on Nearest Neighbors. I’d heard about the algorithm before but the visuals made it far clearer. One of my security courses has a project component at the end of the semester, I might try to implement the algorithm for something down the road.
Wednesday, September 28, 2016: 8:54am – 6:12pm (9.3 hours):
I finished the assignment in my morning course really early so I got into lab really early as well and made some updates to the slides Kaushik and I would present with.
The interface needed some tweaking still, and I wanted an outline of how the pipeline would eventually function. I incorporated the code for generatePrimaryShares.py in runthis.sh as a demo so something would at least happen when you clicked the button. Now that a lot of the main features at least has an outline associated with it, I’ve officially dubbed this as Beta V0.3.
Kaushik and I had a successful presentation, and I made quite a few updates to the interface throughout the day. By the end I had changed the logo, implemented a proof-of-concept for switching between directories, and made a huge number of visual changes. The changes seemed significant enough that I bumped the counter up to Beta V0.3.
Thursday, September 29, 2016:
I missed reading group, it was held at 10:00am instead of it’s usual time. This was a good reason to be added to the WhatsApp group though.
Friday, September 30, 2016: 12:48pm – 4:24pm (3.6 hours):
Kaushik and I reviewed the pipeline and made sure everything was correct in the README, making sure I had a solid grasp on it was also helpful for thinking about implementation again for the interface.
The wrap up the week Phillip gave a practice talk about his work.