Most of the REU students were ready to begin bright and early after everyone got a good night sleep the night before. We spent the morning meeting with our advisers and getting a brief outline for the rest of the semester. I met Dr. Natarajan (henceforth to be called: “Professor”) and he described our project as focusing on the topic that appeared in the latter section of his “Relational Learning for Sustainable Health Paper” (refer to Day 3 report).
The project goes something along these lines:
- Step 0: Compile blogs and information to create a database for drugs and their adverse effects.
- Step 1: Use machine learning to validate that these effects are actually caused by these drugs. (“A Novel Method for Adverse Drug Event Extraction from Text”)
- Step 2: Build in natural language processing to account for the top 50 research papers (bringing in data from abstracts listed on PubMed), in addition to the information from blogs and research notes that can be mined.
- Answer: given a patients medical history, describe the reason for a side effect, or conclude that there is no relation between the experiences and prescriptions.
We spent a few hours meeting people, reading papers, and seeing the lab where we would have meetings.
Dr. Natarajan’s words of wisdom:
- Don’t be late, when you’re late you waste the time of everyone who was on time.
- It’s better to be supremely committed than supremely smart. Research is 1% inspiration and 99% perspiration, if you work hard enough the smartness comes naturally.
- Direct feedback is quickest and most helpful.
At 3:00pm, the REU students and some of the faculty gathered for ProHealth Tea in the Informatics Building for donuts, tea, and the elevator speeches we had been preparing.
“Good afternoon, I’m Alexander Hayes and I study Computer Science at Indiana University. There’s a lot of subjects I’m interested in but I really want to narrow down what I want to study as I consider grad school down the road. Previously I’ve worked with Dr. Kapadia on security research and am I’m really excited work with Dr. Natarajan this summer on machine learning. Our project uses natural language processing to analyze medical data and find relations between drugs and adverse health effects, not only is research important because it can potentially save millions of dollars a year, but also has the potential to save tens of thousands of lives.”
And the winners were chosen:
Afterwards I wound down the day by going through a paper on Inductive Logic Programming and left for the week at 5:00pm to enjoy the three-day weekend.