“It was not a good time to be dropped on my head,” said CS sophomore Christy Warden as she reflected on her change from high school cheerleader to college dancer.
Warden had arrived at Rice intent on studying brain activity and decided to protect her own by discontinuing her role as a flier – the cheerleader thrown into the air during stunts. Instead, she successfully auditioned for the Rice Dance Team as a freshman and became a co-captain as a sophomore. She also leads campus tours for the Office of Admission, works for the McNair Center for Entrepreneurship and Innovation, and is double-majoring in Computer Science and Cognitive Science with a minor in Neuroscience.
Her goal of studying brain activity has remained the same even though her path has shifted slightly. “In high school, I thought I wanted to be a bioengineer,” she said. “Not many people in my area had heard of Rice, but it was on the list of ‘best bioengineering schools’ so I looked into it.”
Arriving for an overnight campus visit, Warden said, “My host was so excited, she told all of her friends that I was coming and was so welcoming to me. I really enjoyed talking to her and her friends about their experiences at Rice and hearing how enthusiastic they were about their classes and residential colleges.”
A similar experience during her OWL-Days visit pushed Rice to the top of Warden’s of list and she accepted the university’s admission offer. Her enthusiasm for Rice remained high, but she was less excited about some of her required courses. “During O-Week when we were registering for classes, I looked at my BioE requirements and saw computer science on the list. I didn’t really like computers and decided to get it over with quickly,” she said.
But COMP 140 surprised her. “We’re immediately assigned to teams in the Introduction to Computational Thinking. I was matched with two other freshmen girls from my residential college and we became really close friends,” said Warden. “Plus, Rebecca Smith was our professor. She’s so talented and became a really big inspiration for me.”
Warden also began realizing the types of brain activity problems she wanted to solve could be approached through the lens of computer science. She said, “I realized I could study machine learning and artificial intelligence and that those areas more closely aligned with what I wanted to be doing for my career.”
After completing a second COMP class, one of her friends from COMP 140 suggested she work as computer science research assistant in Rice’s McNair Center for Entrepreneurship and Innovation. “She was already working there and encouraged me to apply. I wasn’t sure it was a good match for my skills, but it was convenient and it’s actually become a great fit for me,” said Warden.
Her job started with data parsing and web crawling as she helped the policy researchers gather statistical data. “I also do a lot with Twitter crawling and have four bots running on Twitter right now.” Warden said Twitter offers an application program interface (API) that customers can use to mine the social media channel’s database. “You get an access key and a secret code for their database so you can search for specific data on trends and user information.”
Recently, Warden’s supervisor asked if she could work on neural nets, a method of training computers to perform like the neural systems found in brains. Warden said she felt elated at the prospect.
“Because I’m taking a lot of neurosystems classes, I already had in mind that I wanted to work on a computer program inspired by the brain. That is what I had always wanted to do, and I was asked if I could do it!”
The process she is following utilizes previously labeled datasets. Warden said, “The input we provide is a description of a company, and the output we want is an industry category. You train the computer by giving it a dataset that already has the expected output matched to the input. After the computer practices on thousands of previous examples, then your neural net becomes adept at giving you the correct industry classification for a description. Then when you provide new input, you can be confident it will be processed correctly.”
She’s also used existing libraries to calculate the distance between the expected outcome and the output from the neural net during the training phase of the classifier. “I use backpropagation to make adjustments until the error between the result and the correct answer is minimal,” she said.
Warden applied a little backpropagation to her study habits, and offers prospective CS students advice based on her own lessons learned.
“In my third COMP course, I realized I’d developed several unhealthy habits, mostly by trying to work alone,” she said. “I’d open the assigned project, not understand it on the first read through, get intimidated, procrastinate, not go to office hours, then turn in my homework right at the deadline and know it was not my best work.”
She decided to change her style by finding a study buddy. “Find a companion to go through CS with,” she said. “That’s been a big game changer for me this year. I talked with Jeemin Sim – she and I worked well together in a previous lab – and we decided to work on projects together.
“It is so much easier to open up a project when you are sitting down with someone else and going through it together. I understand some parts, she understands other parts, and we talk through different aspects and approaches. It’s not like we open up our code and write the same thing, but it is enormously helpful to have someone to brainstorm with.”
Both Warden and Sim signed up together for COMP 321, a class on computer systems where the instructor encourages students to work with a partner. “It was actually a little scary, asking someone to commit to that kind of grade-influencing partnership,” said Warden. “But Jeemin and I work really well together and even set up a Google Calendar to block out the hours we’ll work together and when we’ll go to study hours.”
Committing that time in advance also helps Warden balance her other activities – work, dance practice and performances, research, and her other homework.
“It is so easy to feel like you are drowning,” she said, “but knowing I will start the project with Jeemin makes it easier to open that big seven-page project spec. Sitting down with someone and reading and explaining the project to each other is a much better way to begin and has had such a positive influence on my attitude towards my homework and my skills as a student.”