Shivani R. 2023 | BASIS Independent Silicon Valley
- Project Title: Evaluating the Effectiveness of Two-Step Transfer-Learning for Emotion Detection Software for Children with Autism
- Basis Independent Advisor: Jon Noble
Through AP Research, I have been investigating the use of two-step transfer-learning as a method to improve the accuracy of emotion detection software for children with autism. The field of artificial intelligence (AI) and human-computer interaction (HCI) has been of great interest to me, particularly in regards to its potential to positively impact individuals with special needs. I embarked on this research project with the understanding that open-source emotion detection software often lacks the performance capabilities to accurately detect emotions in individuals with autism. With this in mind, my objective was to make this technology more accessible to this population by evaluating the effectiveness of two-step transfer-learning. Transfer-learning involves pre-training a model on a large dataset, followed by fine-tuning on a smaller dataset specific to the task at hand; in two-step transfer-learning, this process is done twice. This method is used when there is minimal open-source data for a specific research project. With the help of special education professionals, I have created my own datasets for my research. The potential applications of this technology are vast, including its use as an educational tool to provide real-time feedback to students on their emotional engagement during lessons, and as a tool in behavior therapy to aid in the understanding of emotional states of patients. This can lead to improved communication between caregivers and children with autism, ultimately resulting in more effective treatment. This research contributes to the growing body of literature on the use of AI in assistive technologies and the potential for its use in the field of autism.