Leanna Fan, a 14-year-old middle school student from San Diego, California, won this year’s 3M’s Young Scientist Challenge top prize! Her low-cost Finsen Headphones, which use machine learning and phototherapy, treat mid-ear infections in children by emitting blue light at the target, helping to prevent up to 60% of hearing loss. Leanne worked with 3M’s material laboratory research assistant Dr. Ross Behiling over the summer, making her concept into a prototype. All nine finalists who participated in this contest presented their innovations at the 3M headquarters in St. Paul, Minnesota.
Approximately 700 million people worldwide are diagnosed with mid-ear infections, with close to 21,000 deaths per year. Most cases involve children living in underprivileged countries. Those without medical access or healthcare have a difficult time receiving diagnosis and treatment. Leanne’s headphones could solve this issue due to their affordability and ability to detect and treat mid-ear infections. Leanne plans to use her prize money to begin patenting the prototype.
Here are a few finalists from the 3M Young Scientist Challenge:
Amritha Praveen, a 13-year-old junior high school student from Buffalo Grove, Illinois, developed the ASD screener, a tool that uses machine learning to predict autism spectrum disorder based on 16S ribosomal RNA sequencing of the human gut microbiome. She also downsized the features via a reduction algorithm because the microbiome data is highly dimensional, zero-inflated, non-linear, and suffers from cursive dimensionality.
She developed this program in Python with scikit-learn machine learning libraries. Then, k-fold cross-validation evaluated the model, and grid search adjusted the model’s hyperparameters. Her final model has a 92% ASD prediction success rate, an improvement over the initial model trained with the original features.
Harini Venkatesh, a 14-year-old student from Brentwood, New Hemisphere, created The Comptometrist, a prototype that quickly calculates the myopic power in a patient’s eyes. To develop the prototype, he wrote a program that calculates the aspect ratio of an ellipse. This program uses the OpenCV Python library to find an ellipse in an input image and calculate the quotient of the largest and smallest diameters of the ellipse.
Next, he collected 40 eye images from a local optometrist who sent scans and full prescriptions. Afterward, he discovered the open-source Foveal Avascular Zone Image Database on the Open ECPSR website, which provided access to myopic and non-myopic eyes. He expanded his dataset to use 100 images for model training and 20 for testing purposes. As a result, it returned an average of 0.5 diopters, nearing the acceptable error in a manually generated prescription.
Asvini Thivakaran, a 13-year-old middle school student from Round Rock, Texas, discovered a unique approach that enables electric vehicles to charge while driving. She created a model vehicle using a toy car and a piece of material on the tire with a photo rectifier and capacitor. The rectifier converts the piezo to generate AC to DC, and the capacitor stores the energy. She drove the vehicle to various distances and used a multimeter to measure the voltage across the capacitor. Afterward, she calculated the energy stored in the capacitor and prorated it to an actual vehicle. According to her findings, an electric vehicle with a piece of material on all four wheels that travels one mile at 5 miles per hour generates 0.0693 electric power. This allows the vehicle to travel 0.2 more miles, a 20% increase. This solution could help solve the global warming crisis.
Have a story tip? Message me at: http://twitter.com/Cabe_Atwell
