Tejasvi Kothapalli

I am a first-year Vision Science PhD student at UC Berkeley. I am in the the Active Vision and Neural Computation Lab and advised by Professor Jacob Yates. I am working at the intersection of computational neuroscience and computer vision. I am generously funded by the CIVO Fellowship.

I was previously a student research assistant jointly at UC Berkeley's Clinical Research Center (CRC) and the International Computer Science Institute (ICSI). There I worked with Prof. Meng Lin, Prof. Stella Yu, and Peter Wang.

I graduated from UC Berkeley in May 2022 with a Bachelor of Science degree in Electrical Engineering & Computer Science.

Click for more about me!

email: tejasvi.kothapalli [at sign] berkeley.edu

CV  /  Transcript  /  Google Scholar

profile photo

Harsh, Avinash, Me, Andi

News
  • [08/2023] Started PhD in Vision Science at UC Berkeley

  • [08/2023] Named CIVO Fellow by the UC Berkeley Center for Innovation in Vision and Optics

  • [10/2022] Paper accepted at NeurIPS 2022, Medical Imaging Workshop

  • [05/2022] Received UC Berkeley EECS B.S. degree
Research
A Machine Learning Approach to Predicting Dry Eye-Related Signs, Symptoms and Diagnoses from Meibography Images

We propose a machine learning approach that predicts dry eye related signs, symptoms and diagnoses from subject data.

Tracking the Dynamics of the Tear Film Lipid Layer
Tejasvi Kothapalli, Charlie Shou, Jennifer Ding, Peter Wang, Andrew D. Graham, Tatyana Svitova, Stella X. Yu, Meng C. Lin
NeurIPS Workshop, 2022
project page / paper / poster / slides / arXiv

We use computer vision to track the tear film lipid layer spread.

Senior Honors Thesis: Studying Dry Eye Syndrome with Machine Learning
Tejasvi Kothapalli
Senior Honors Thesis, 2022
thesis

Studied Dry Eye Disease diagnosis with machine learning in two regimes: (1) patient tabular data and (2) ocular surface recordings. Advised by Prof. Stella Yu and Prof. Meng Lin. Earned an A+ :). Also, huge thanks to Peter Wang for all the mentorship and guidance.

Tear Aqueous Production Rate Clinicial Tool
Tejasvi Kothapalli, Jennifer Ding, Andrew D. Graham, Meng C. Lin
2021
project page / clinical tool

This tool (currently being used in clinic) computes the tear aqueous production rate following data collection from a Schirmer test.

Saving Energy in Homes Using Wi-Fi Device Usage Patterns
Tejasvi Kothapalli
International Journal of Energy Optimization and Engineering (IJEOE), 2018
paper / publication site

In order to reduce energy wastage in homes, this work detects the presence of residents through Wi-Fi device usage and tries to automatically turn off unused electronic appliances.

Controlling Tensegrity Robots through Evolution using Friction based Actuation
Tejasvi Kothapalli, Adrian Agogino
NASA Technical Reports, 2017
paper / publication site

This study investigates a control pattern for successful locomotion in tensegrity robots through an evolutionary algorithm.

Notable Coursework
CS 280 Final Report: Analyzing the Spread of the Tear Film Lipid Layer with Patch Tracking
Tejasvi Kothapalli
2022
Report / Slides

This is a very preliminary version of Tracking the Dynamics of the Tear Film Lipid Layer.

CS 285 Final Report: Employing Recurrent Policies for Meta-Learning to Play Sonic The Hedgehog™
Brian Zhu*, Tejasvi Kothapalli*
2021
Report

We approached Open AI's Retro contest with two new approaches. We first explore meta-learning with recurrent policies. Second, we explored the use of enhanced exploration with Random Network Distillation.

CS 190-80 Final Report: Converting Equation Image into Latex Markup
Brian Zhu*, Tejasvi Kothapalli*
2021
Report

We created a model that converts images of LaTeX into its original markup expression.

CS 182 Final Report: Understanding How Different Segments of a Yelp Review Affect Overall Star Rating Prediction with BERT
Tejasvi Kothapalli*, Brian Zhu*, Joe Zou*
2021
Report

The intention of our technique was to understand how different segments in a review affect the overall prediction through sliding window methods. We hope that our contribution can help people better understand how these models make decisions.

Teaching

After graduating from Berkeley, I have taught the Inpsirit AI Scholars curriculum to high students in-person at Khan Lab School, Bellarmine College Preparatory, and Bentley School. Teaching at the Khan Lab School in July 2022 with my good friend Sohum Hulyalkar was a truly rewarding experience.


Huge thanks to Jon Barron for website template!