RESEARCH
My research reflects a deep interest in learning and connecting diverse fields from the arts to STEM. I enjoy integrating the things I love with scientific inquiry, whether through data-driven environmental work or technology's creative applications. These projects represent an ongoing commitment to exploration, and I hope to continue expanding this work through further study and collaboration.
Role of Emotion In Art
Throughout, I have maintained a strong interest in the Arts. With the emergence of AI-generated art and imagery, this interest evolved into a curiosity about how machine-generated art labels align with human emotional responses to similar visuals. Understanding emotional diversity in artistic expression offers valuable implications for therapeutic and creative applications. Motivated by these possibilities and supported by expert guidance, I chose to explore the intersection of AI-generated text, artistic form, and human emotion as a research focus.
Underwater Pollution Detection
Utilizing YOLO and Computer Vision for Real-Time Detection of Marine Plastic Pollution
Abstract
Underwater plastic pollution poses significant risks to marine ecosystems and human health by threatening biodiversity and contaminating food sources. Precise detection is essential for implementing timely corrective actions to conserve marine life and protect the environment. This study explores the application of YOLO-based deep learning models, particularly YOLOv8 and YOLOv10, to effectively detect and classify submerged plastic debris. By leveraging these models, the project can significantly enhance existing pollution control strategies through scalable real-time monitoring.
With YOLOv8 achieving a higher mean average precision (mAP) of 0.856 over YOLOv10's 0.843, the research underscores the importance of accurate detection for effective conservation efforts. The findings highlight the potential for these models to provide automated tracking systems that support ongoing marine pollution management. Looking ahead, the project aims to develop a prototype capable of handling real-time data, facilitating improvements in pollution response and prevention, ultimately safeguarding our oceans and their delicate ecosystems.
Submissions
Paper completed in September 2025
Submitted to October 2025
Applications
I am currently working on the Conrad Challenge, where our project involves developing a WasteShark-based hardware concept integrated with the machine learning model from my research. The goal is to identify highly polluted waterways using satellite imagery and deploy the device to detect, make decisions, and collect the most harmful debris while gathering environmental data.
Separately, I have also developed an app that allows users to upload underwater images for automated plastic classification, supporting data collection efforts for marine conservation.
Download the app here
2024-2025
Role Of Emotion In Art
Cross-Modal Semantic Coherence in Art: The Role of Emotion and Style in Human–AI Understanding
Abstract
This study examined how the type of emotion, art style, and their interaction affect the coherence between human emotional descriptions and AI interpretations. A combined dataset was created using ArtEmis art metadata and image-text data from LAION 5B. A multimodal framework was built to analyse emotional and semantic alignment using transformer-based models that represent image and text features in a shared meaning space. Cosine similarity helped measure the degree to which visual and textual emotions aligned.
Statistical tests such as ANOVA, Kruskal-Wallis, and entropy measures were used to assess the relationship between emotion and style. Random Forest and AutoGluon models further validated the hypotheses and tested predictive accuracy. Results showed that positive emotions, such as awe, amusement, and contentment, had a stronger alignment between human and AI interpretations than negative emotions. Representational art styles such as Impressionism and Romanticism displayed better semantic consistency than abstract styles, which showed higher emotional diversity but lower alignment.
Overall, the analysis revealed that emotion and style each played an independent role in improving cross-modal coherence; however, their interaction did not significantly alter the outcomes. The study also proposed a sequential attention-based model that better captures links between emotion and meaning. The findings bridge computational aesthetics and affective computing, demonstrating how both emotional tone and visual realism influence the way art communicates—making them particularly useful for generative AI, digital curation, and art therapy applications.
Submissions
Paper completed in September 2024
Submitted to Journal Of Emerging Investigators, October 2025
Context
I was mentored by Dr. Shailendra Kadre at Christ University, Delhi. This was my first opportunity to engage in advanced, university-level research, building on ideas from my earlier paper on generative AI in the art industry. I aimed to take those concepts further through more rigorous academic inquiry. Looking ahead, I hope to pursue similar high-level research in marine biology, particularly investigating coral resilience and the effects of climate change on marine ecosystems.
2023 - 2024
Generative AI in Art
How Generative AI Is Reshaping Artistic Creation, Professional Practice, and the Structure of the Art Industry
Abstract
This paper examines the influence of generative artificial intelligence on the contemporary art industry, with a focus on how AI challenges traditional ideas of creativity, authorship, and artistic value. It provides an overview of the fundamentals of generative AI, including the machine learning and deep learning models that enable it to produce visual content. The paper compares human and machine creativity, emphasizing the emotional, intentional, and experiential elements unique to human artists. To understand real-world perspectives, interviews were conducted with practicing artists working across different mediums. Their views highlight both the potential of generative AI as a creative tool and the concerns surrounding originality, copyright, and the democratization of art production. While generative AI introduces new opportunities for artistic experimentation and accessibility, it also raises ethical and professional challenges that the industry must address. Overall, the findings suggest that generative AI will not replace human artists but will continue to reshape artistic practice and expand the definition of art in the future.
Context
This paper was my first independent research project, completed in the summer after Grade 9 under the mentorship of Professor Phanish Puranam, Professor of Strategy at INSEAD. At the time, I was deeply interested in art, and with generative AI rapidly emerging, I wanted to explore the intersection between the two. I also had the opportunity to interview Southeast Asian artists to understand their perspectives on how these technologies may shape the future of the art industry.
As an early research experience, this project was not developed with the intention of publication, but it played a meaningful role in building my foundational research skills. If you are interested in reading the full paper, feel free to email me and I would be happy to share the PDF.
2023 Summer