Stanford University
2023
Researcher
About the project
I have had the extraordinary opportunity to secure a selective research position at Stanford University's Computer Science Department and in collaboration with NYU Stern.
I competed alongside thousands of seniors and freshman undergraduate students. I am the only person from India to receive this position this year. Every year Stanford takes in 60 interns to assist each department, where the majority are college students, and high schoolers residing in California. However, being part of one of the only 5 internationals and a high school student speaks to my academic abilities.
For the research project itself, we aimed to replicate human behaviour and human decisions. Our research aims to train two models, Convolutional Neural Networks (CNNs) and fully connected Neural Networks (NN), that can predict the next action of a user given the current state. Predicting the user's next action is what we call “cloning”. We have achieved a 93% accuracy rate in replicating human behaviour using these two models.
I was invited back for the second phase of this project to work with NYU, specifically Professor Zhou and his team.
My role
Training
Engineering and Modelling
Evaluation
I underwent lots of training with professors across Stanford University to learn new things in the fields of Computer Science, Economics, Artificial Intelligence, and Public Policy.
We had to concatenate all the data and create heuristics, which are basically AI algorithms that generate data like a human to give us data for the neural networks to be trained on.
We programmed the CNN. We implemented multiple layers for full accuracy and then image-trained it using tensors and directions.
Coded CNNs to print the loss and accuracy after each epoch of the training dataset. After which using this data we had to plot this on Matplotlib to visualise it.
01
My Learnings
Working on a project that involves both computer science and behavioural insights fostered collaboration across diverse disciplines. I learned to bridge the gap between technical aspects like machine learning and the human behaviours I’m studying. This interdisciplinary exposure hones my ability to communicate effectively and collaborate with professionals from various backgrounds.
02
My Learnings
This project allowed me to witness how machine learning can be applied to solve real-world challenges. Whether it's optimising energy consumption, predicting financial trends, or informing public policy, I saw firsthand how technology can have a tangible impact on various domains. This understanding enhanced my perspective on the role of technology in creating positive change. Behaviour logging involves collecting data on how occupants interact with the building environment. This could include factors like lighting usage, temperature preferences, appliance usage, and occupancy patterns. By monitoring and analysing these behaviours, insights can be gained into how energy is being used and potentially wasted. Learning from this can allow houses/buildings to replicate human behaviour and therefore exponentially decrease the energy usage we have.
03
My Learnings
Under the mentorship of Professors Zhou, Pilanci, and Auclert, I had the privilege of delving into diverse academic realms. With Professor Zhou at New York University (NYU), I immersed myself in machine learning, gaining insights into both theoretical foundations and practical applications. Professor Pilanci's guidance in electrical engineering provided a comprehensive understanding of complex concepts and real-world applications, fostering innovation and problem-solving. Under Professor Auclert at Stanford University's Economics Department, I honed critical thinking skills and developed a passion for socio-economic analysis and policy-making. These experiences have expanded my knowledge and shaped my academic and professional pursuits, inspiring a lifelong commitment to innovation and meaningful contributions to various fields.