Shlok Bhattacharya, a senior at Chagrin Falls High School, immersed himself in the research process through his enrollment in the schools AP Capstone program during his sophomore and junior years.
As a junior, Shlok’s research aligned with his passion in the field of machine learning and his interest in making a positive impact in society. With the growing popularity of social media, he realized that offensive language has been able to proliferate at an unprecedented rate. As a result, Shlok decided to research hate speech detection. Specifically, he was driven to answer the question, “To what extent does the creation of a culturally-weighted training dataset impact the ability of a support vector machine model to mitigate cultural bias in the context of hate speech detection?” His research, entitled “Battling cultural bias within hate speech detection: An experimental correlation analysis,” explored this question.
Shlok’s work filled gaps in existing research, as it explored a new approach to mitigating cultural bias in hate speech detection, namely Categorical Weighted Training (CaWT). Previous studies hinted at the potential effectiveness of methods like CaWT, but none had explicitly applied it to the issue of cultural bias in machine learning models for hate speech detection. Shlok indicated that while hate speech detection systems are vital in today's digital landscape, they often struggle to discern hate speech from clean speech within different cultural contexts, leading to cultural bias. This study was intended to fill the gap by investigating the correlation between the cultural weight of a training dataset and the model's performance, ultimately showing that a lower cultural weight is ideal for mitigating bias within the research constraints.
Prior to November 2022 and the widespread release of ChatGPT and other AI tools, Shlok was driven by the question, “Could you teach a computer like a student?” This question became his obsession after stumbling upon the concept of machine learning during the 2021-22 school year. Over the years, Shlok had spent time tutoring students in all grade levels, so the idea of teaching a machine was very interesting to Shlok. Through his research and related studies, he decided to join the Association for Computational Linguistics Anthology. This enabled him to explore various applications of machine learning and natural language processing, actively seeking opportunities to apply his newfound knowledge. Once he identified the societal issue of hate speech, he realized that he could bridge these two areas of interest and embark upon his research journey.
After completing his research, Shlok felt a strong drive to share its insights regarding hate speech and online abuse mitigation through machine learning and artificial intelligence. He sought advice from Dr. Sandra Kübler at Indiana University, Bloomington, who recommended professional conferences and reputable science journals as communication vehicles. From her recommendations, he identified the Journal of Emerging Investigators (JEI), a peer-reviewed scientific journal as a publication which might be most aligned to work like his.
During a summer program at Stanford University, Shlok had the privilege of sharing his research with fellow undergraduates and high school students he had met in my classes. He was excited by their enthusiasm in learning about his work and suggested it had potential for publication. This feedback encouraged Shlok to collaborate with Mrs. Jean Kanzinger, his AP Research teacher, leveraging her valuable contributions, to submit their research to JEI. This manuscript has already been accepted, with an expected publication date in JEI by the end of November 2023, pending final preparations.
Shlok’s research experiences built his interest and foundations in machine learning. He plans to use these essential skills while furthering his studies in the field of computer science, with a specific focus on artificial intelligence and machine learning. In particular, he plans to further study natural language processing and computational linguistics.
In the future, Shlok hopes to explore the exciting realm of quantum computing and actively contribute to its continued development. The foundation in machine learning that he gained through his research will undoubtedly be a valuable asset as he seeks to grow his knowledge and make meaningful contributions to the field of computer science and artificial intelligence. “Shlok’s enthusiasm for learning and interest in contributing to the field and making a positive difference in society reflects his commitment to grow and thrive,” said Principal Amanda Rassi, Chagrin Falls High School. Director of Curriculum Becky Quinn added, “The AP Capstone program affords students like Shlok with the opportunity to grow their research skills while identifying topics and questions that reflect their unique passions and interests. The student agency afforded to students through these courses has the potential to yield amazing future impact - for our students, our community, and the world.”