THURJ Research Competition
Led by Harvard students to promote research experience and skills for high school students.
Register by March 31st 11:59 PM PST.
-
Application for the Spring 2025 competition is OPEN.
-
Open to all MS/HS students!
-
Register here: https://tinyurl.com/thurjcompforms25
-
Work in teams of up to four! Receive mentorship from experienced Harvard researchers and peer reviewers!
-
Sign up for our March 24st Info Session at 9:30PM EST: https://forms.gle/aaSLZz62Jtqsg6MC8
FALL 2024 Awards
​Congratulations to our inaugural high school research competition winners!
​
First Place: Joe Liang
Second Place: Samantha Narchetty
Third Place: Anusha Kumar
Fourth Place: Aami Dahiya
Fifth Place: Sudarshan Mahesh
Award Winners Features:
1st Place
Joe Liang
Summary of Work
The double pendulum has long been a favorite tool in classrooms to illustrate unpredictable behavior and chaos. This paper uses the Lyapunov exponent, a quantitative indicator of chaos, calculated at different initial conditions, to identify periodic trajectories of the double pendulum. The findings suggest that even in highly sensitive systems, periodic orbits can emerge, challenging the notion that chaos is entirely unpredictable and offering a method applicable to other chaotic systems.
3rd Place
Anusha Kumar
Summary of Work
My research focuses on developing a machine learning model to predict the probability that genes are associated to disease based on connectivity to other genes. We generated 128-dimensional node embeddings based on connectivity to train the classifiers. Our machine learning models identified misclassified genes in existing databases and reclassified them, generating 1,019 consensus candidate disease genes for further biological research.
5th Place
Sudarshan Mahesh
Summary of Work
This work introduces a cost-effective, gesture-controlled 6-DOF robotic arm designed to enhance independence for elderly individuals with reduced mobility, focusing on practical applications for daily tasks. By integrating Computer Vision with OpenCV and Arduino-based hardware, the system utilizes facial gesture recognition to control the robotic arm's pick-and-place operations, offering a simple yet effective interface for users. The design employs real-time image processing, low-latency serial communication, and optimized servo actuation to ensure adaptive user interaction.
2nd Place
Samantha Narchetty
Summary of Work
Robotic swarms hold the potential to solve complex problems more efficiently with promising applications in Search and Rescue (SaR) operations. This paper employs a more realistic evaluation framework using the PyBullet simulation software to test the performance of four search approaches based on swarm intelligence algorithms: Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Glowworm Swarm Optimization (GSO), and Rapidly Exploring Random Trees (RRT) in search and rescue scenarios. These algorithms were tested across various swarm sizes and analyzed to determine the best algorithmic search approach depending on resource availability.
4th Place
Aami Dahiya
Summary of Work
This paper explores how childcare availability affects women’s workforce participation in India, focusing on Mobile Creches, an NGO that provides childcare centers at construction sites. Through interviews with working mothers, the study highlights how access to creches improves work-life balance, increases income, and enhances children's well-being. Despite these benefits, challenges like affordability, accessibility, and gender biases persist, highlighting the need for expanded childcare services to support women’s economic empowerment.
Best Manuscript for Underrepresented Researchers
Arkangel Cordero
Summary of Work
The use of observational data in mediation analysis to determine whether an observed relationship between an independent and dependent variable causally operates through a mediator has been a long-standing challenge for researchers. Since mediators are endogenous, standard regression approaches can produce misleading estimates, making it difficult to separate true causal pathways from confounding correlations. This paper repurposes the instrumental variable method, using properly lagged levels of the mediator as instruments for its first-difference transformation, mitigating bias and ultimately allowing researchers to apply existing under-identification and weak-identification tests, providing a practical framework for mediation analysis.
Hear from Past Participants
"The competition definitely helped me improve my scientific ability as a researcher. Prior to this competition, I had never worked on a research paper, and I learned so much throughout this process. Most importantly, the competition has served to help me enhance my writing skills, as well as improve my ability to analyze data."
-Fall 2024 Competition Participant
SPRING 2024 Awards
Congratulations to our inaugural high school research competition winners!
​
First Place: Disha Gupta
Second Place: Aditya Shivakumar
Third Place: Churan Xu
Fourth Place: Nathan Hu
Fifth Place: Saanvi Bhargava
Award Winners Features:
1st Place
Disha Gupta
Summary of Work
There is a huge and growing disparity in global breast cancer care: mortality rates stand at 7% in the US, compared to 27% in India and 60% in Africa. This is primarily because of the delay in diagnosis and late-stage intervention in developing countries. This study focuses on reducing global breast cancer mortality and cost of care by enabling accurate, early, low-cost diagnosis using the existing ultrasound infrastructure and medical expertise in underserved communities and applying generative AI and a hybrid deep-learning architecture.
3rd Place
Churan Xu
Summary of Work
Lila Abu-Lughod and Saba Mahmood are pioneering anthropologists who ground their ethnography in underprivileged communities across the Middle East. In this essay, I explore how they utilize desire as an anchoring concept to expand on the meaning of agency and freedom amongst the backdrop of an overarching Islamic patriarchy, highlighting how their work aims to defy the top-bottom conclusions of the Western gaze and Orientalism with concrete storytelling.
5th Place
Saanvi Bhargava
Summary of Work
This work presents a Machine Learning model for automatically grading vocal music recordings cumulatively on pitch and rhythm, given a reference piece of music. In addition, a correspondence algorithm is also presented, which is used to map notes in the performance sample to the notes in the reference sample and subsequently give granular feedback on a note level, such as pitch mismatch, notes started late, missed notes, and many other factors.
2nd Place
Aditya Shivakumar
Summary of Work
Cepheid variable stars act as "cosmic candles," offering standard measurements of cosmic distances due to their intrinsic Period-Luminosity (P-L) relationship. Therefore, accurate classification into Delta Cepheids (DCEP), Type II Cepheids (T2CEP), and Anomalous Cepheids (ACEP) is crucial for refining the cosmic distance ladder. To address this classification problem, we employ Recurrent Neural Network (RNN) autoencoders to extract latent features from phase-folded photometric light curves and introduce a novel feature, L2DCEP, based on the Period-Wesenheit relationship.
4th Place
Nathan Hu
Summary of Work
Despite growing public concerns about high school students’ mental health, research into how kindness-based interventions affect their subjective well-being remains scant. This study addresses three research gaps in kindness-based interventions: adolescent-peer leadership throughout the experiment, administration outside laboratory or classroom settings, and applicability to boarding high school students. This peer-led, simple, and cost-effective kindness-based intervention can be readily replicated by student leaders in day and boarding high schools on a sustainable and scalable basis, bolstering subjective well-being at both individual and school community levels.
Best Manuscript for Underrepresented Researchers
Lucia Martinez-Pelaez, Meredith Ho, Richard Cavaliere-Mazziotta, Jacob Parker
Summary of Work
Our paper dives deeply into the neurological as well as social effects of pornography consumption on adolescents. Looking at the issue from a neurological standpoint, we highlight the desensitization and dysfunction that is created due to avid porn consumption in this age group. Looking at this subject from a social stance, pornography negatively affects adolescents' relationships and behaviors due to the unrealistic portrayal of sex it creates.