La Cañada High School sophomore Jonah Garland said the inspiration for his first-place science fair project came from machine learning, artificial intelligence and, maybe most important, breast cancer.
At LCHS’ Science Fair last Friday evening, Garland presented a different way to break down data on breast cancer mammogram compilations, classifying images into different categories such as negative, benign or malignant mass, using computer-aided diagnoses to “increase speed, accuracy and precision.”
Garland, who will compete at the Los Angeles County Science and Engineering Fair in late March at the Pasadena Convention Center, took home a blue ribbon and a certificate last week for his project “Applying Neural Networks to Breast Cancer Classifications.”
The top three winners at the LCHS show move on to the L.A. County competition. Ten projects put forth by 16 students competed in the event sponsored by the LCUSD science enrichment programs. There were seven judges, including three who work at NASA Jet Propulsion Laboratory, a senior packaging engineer from Nestle, a family and child psychotherapist. A small but enthusiastic group of parents and students also attended the competition at the LCHS cafeteria.
“It’s a very widespread problem that applies to all women,” Garland, 16, said of the threat of breast cancer. “Everyone gets tested. We have had incidences of breast cancer in my family, so it connected to me on a personal level in that way. It’s just a widespread problem that really needs an answer.”
Garland, who lives in LCF, described his project as depicting a process of machine learning that is leading the way for computer-aided diagnostics.
The exhibit featured a large white board featuring a definition of machine learning as “an application of artificial intelligence that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.”
Another definition included one for artificial neural networks, described as being “able to perform many of the same pattern recognition tasks as humans, with increased accuracy, speed and precision.”
According to Garland’s background on the project, “Digital Reasoning released a computer neural network with 160 billion neurons, 60 billion more neurons than the average human brain.” Digital Reasoning, based in Franklin, Tennessee, offers cognitive computing services to intelligence agencies, financial institutions and health-care organizations, according to its website.
“What we can do is we can take a neural network — which is an artificial neural network on my computer — which is going to classify breast cancer mammograms and can really lead to a bypass of the biopsy,” Garland said. “It applies to a lot of people and it’s really just a mathematical model that does a lot of very complex things. Most importantly, it’s a branch of artificial intelligence that learns from itself.”
Garland said he is planning to create a research program over the summer about machine learning and medical imaging, but if that doesn’t come to fruition he will continue with his breast cancer project over the summer.
LCHS interim Principal Jim Cartnal said before the winners were announced that it was a victory for the students just to participate at the fair.
“Seriously, you’re awesome,” Cartnal said to the entrants after judging finished. “The judges thought very highly of all of your projects.”
Yasmine Kaki, whose project was titled “Testing Memory,” was awarded third place. She said afterward she tested how short-term memory was affected by age, gender, education level and exercise.
“I’m very interested in psychology and the way that the brain works,” the 16-year-old sophomore said. “I always think it’s interesting how older people are like, ‘Oh, my memory used to be so good when I was younger.’ I wanted to test if age actually did affect memory.”
Kaki, who assembled a test group of strangers at a nearby Ralphs, said she found that children, around the ages of 3 to 5, had the worst memory. The best, or fastest, memory times were among people ages 13 to 18, she said. People between the ages of 19 and 49 had the second-best memories, but recall became worse as they got older.
Additionally, Kaki said, women tested better than men in most of the age groups. Those with higher education levels also had better times. With people who exercised, however, she found that those who are 19 to 49 years old and active had worse reaction times. But those aged 50 to 64 who exercised more had better reaction times.
“I think it would be really interesting to increase the amount of people, because I only did 98 people and it would be interesting to see the category of exercise more,” Kaki said. “That one was really tricky, especially because in some of the age groups we didn’t get that many people.”
Second place went to Erin Buchanan, 16, an LCHS junior. In her project “Shed Light on Global Warming,” she compared her data (which included testing the reflection of all the colors on the visible light spectrum including black and white) to data collected in Greenland about their ice caps melting. It matched.
“I thought it was incredible we could quantify something that seems so unquantifiable,” Buchanan said.
Amy Nespor, the LCUSD’s community science liaison and a judge for the science fair, said she felt the projects were generally strong and noted the students’ initiative and enterprise.
“As always, I’m impressed with the kids who are able to do these, because there’s no class, no mentors and there are no teachers,” Nespor said.