My research focused on developing a mobile app with a facial emotion recognition system to treat autism spectrum disorder (ASD). ASD affects 1/40 children in the US. Currently, making one-to-one individualized therapy more accessible to patients with ASD poses a challenge to the healthcare system. An artificial intelligence (AI) based mobile intervention that delivers social therapy is a potential solution. However, existing emotion-recognition platforms have shown limited performance on children with ASD. A novel AI-based emotion classifier was trained with images of children with ASD crowdsourced from a mobile charades-style game called Guess What?. The resulting classifier outperformed existing emotion-recognition platforms for the population of interest. This study supports a new strategy to develop at-home precision therapy for children with ASD.
Jessica Saal Fellowship Award Recipient
Shout out to my amazing mentors!
Stanford Institutes of Medicine Summer Research Program (SIMR) Bioinformatics Group 2019
In our project, we simulated Blackjack games using Python in order to determine what the best strategy truly is. We tested for the best value for the player to stand at (1-21) in order to have the highest probability of winning, as well as how using card counting affects the amount of money a player wins. We also included various options such as adding more players and decks. Our results show that 14 is the best value to stand at whether you are using or not using card counting, but the player only wins roughly 42.5% of the time. This strategy that we have discovered could help to optimize a player’s chances of winning. However, we have found that the player will ultimately always lose money against the dealer when playing a large amount of games and betting without considering the dealer’s shown card.