Andrew Balch

Andrew Balch

he/him/his

Computer Science student interested in tackling social problems

Posts by Collection

papers

Junior Research Project: Automated Induction Robotics

Published:

For my junior research project at the Governor’s School, I designed and built a platform for warehouse robotics (like those used by Amazon) that were powered through wireless induction. Everything I did in this project, from programming and cloud infrastructure to computer-aided design and manufacture, was entirely self-taught. It placed first at the Tidewater Science and Engineering Fair and was accepted to the International Science Fair, but this was canceled due to COVID. For more details, check out my post here!

Download Paper

Governor’s School Mentorship: Transformer-Based Architecture for Android Malware Classification and TTP Prediction

Published:

In my senior year of high school at the Governor’s School for Science and Technology (a public, dual-enrollment program), I conducted a deep learning research project for Android malware detection under the mentorship of Wes Jordan at MITRE. I used the TensorFlow and data science skills I picked up during COVID to develop a framework that generalized the problem to an NLP task via tokenization and applied a transformer architecture. Although the computational cost far exceed anything I had access to, I still learned a lot!

Download Paper | Source Code

Capstone Technical Report: Data-Driven Event Sequence Visualization of Rectal Cancer Outcomes

Published:

For my Bachelor’s of Science capstone project, I continued work I began in 2020 with Emory University modeling a unique dataset of rectal cancer treatment paths. I cleaned the hand-entered data and developed a novel, data-driven visualization approach that integrates feature selection techniques. Click the link to also see my poster, which was presented at the 2024 UVA Engineering Research Expo!

Download Paper | Source Code

Why Algorithms Remain Unjust: Power Structures Surrounding Algorithmic Inequality

Published in arXiv, 2024

Inspired by the sociologist Erik Olin Wright, I conducted an analysis of the power structures that interact with and influence the ways algorithms impact society to perpetuate injustice. Pending submission to alt.CHI.

Recommended citation: Balch, Andrew. (2024). "Why Algorithms Remain Unjust: Power Structures Surrounding Algorithmic Inequality." arXiv preprint arXiv:2405.18461.
Download Paper

Towards an Accessible, Noninvasive Micronutrient Status Assessment Method: A Comprehensive Review of Existing Techniques

Published in arXiv, 2024

A culmination of 2 years of interdisciplinary research at the UVA Human-AI Technology Lab that I have led since my first year. This survey paper examines how the global issue of micronutrient malnutrition has been addressed, exhaustively bridging the clinical and technological aspects of micronutrient status assessment to provide insights and recommendations for future work to effectively address this problem. Submitted to ACM Transactions on Computing for Healthcare.

Recommended citation: Balch, Andrew, Cardei, Maria A., Kranz, Sibylle, Doryab, Afsaneh. (2024). "Towards an Accessible, Noninvasive Micronutrient Status Assessment Method: A Comprehensive Review of Existing Techniques." arXiv preprint arXiv:2408.11877.
Download Paper

Towards Smartphone-Based Monitoring of Micronutrient Status

Published in ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (BCB ’24), 2024

I was granted another opportunity to present my work on smartphone-based spectrophotometry for micronutrient assessment at ACM BCB 2024! Although I didn’t get to go (the dates overlapped with my MobiCom demo), it is still an honor to have my work represented. For more details on the project, check out our full paper here.

Recommended citation: Andrew Balch and Afsaneh Doryab. 2024. Towards Smartphone-Based Monitoring of Micronutrient Status. In 15th ACM International Conference on Bioinformatics, Computational Biology and Health Informatics (BCB ’24), November 22–25, 2024, Shenzhen, China. ACM, New York, NY, USA, 2 pages. https://doi.org/10.1145/3698587.3701418
Download Paper

Exploring Smartphone-based Spectrophotometry for Nutrient Identification and Quantification

Published in arXiv, 2024

During my time as a UVA Dean’s Research Fellow, I utilized my physical prototyping skills to refine an existing design for a smartphone-based spectrophotometer. I built a custom image analysis pipeline to extract quantitative absorbance spectra data using just a smartphone camera and conducted several experiments to fine-tune the performance of the device on micronutrient samples. I demo’d our approach at MobiCom 2024, and the project was featured in a news article by UVA Today! Submitted to the 2025 IEEE International Conference on Pervasive Computing and Communications.

Recommended citation: Balch, Andrew, Cardei, Maria A., Doryab, Afsaneh. (2024). "Exploring Smartphone-based Spectrophotometry for Nutrient Identification and Quantification." arXiv preprint arXiv:2410.11027.
Download Paper | Source Code

Feasibility of Smartphones for Accessible, Noninvasive Micronutrient Assessment

Published in ACM Conference on Mobile Computing and Networking (MobiCom ’24), 2024

In November 2024, I had the opportunity to demo how a smartphone-based spectrophotometer can be used to quantify micronutrients (vitamin B12) to conference attendees at MobiCom 2024 in Washington D.C.! This was a great exercise of my public speaking and presentation skills, and I got to meet lots of other researchers and learn about their work. I also competed in the Student Research Competition, where I won 3rd place in the undergraduate category! For more details on the project, check out our full paper
.

Recommended citation: Andrew Balch and Afsaneh Doryab. 2024. Feasibility of Smartphones for Accessible, Noninvasive Micronutrient Assessment. In The 30th Annual International Conference on Mobile Computing and Networking (ACM MobiCom ’24), November 18–22, 2024, Washington D.C., DC, USA. ACM, New York, NY, USA, 3 pages. https://doi.org/10.1145/3636534.3698861
Download Paper

poetry

projects

BetterSankey

The technical contribution of my capstone research project, BetterSankey is a python utility that extends the functionality of the Plot.ly implementation of Sankey diagrams to be more data-driven and integrate common data science methods (statistical significance testing, feature selection, and model-building). It was developed for the use case of rectal cancer treatment modeling in close collaboration with the Chief of Colorectal Surgery at Emory Cancer Center. Click the link for more details!

Smartphone-Based Spectrophotometer for Vitamin B12 Quantification

This project was demo’d at MobiCom 2024, and we have a full paper currently under review at PerCom!
As part of my research project at the UVA HAI Lab exploring accessible, noninvasive methods of micronutrient status assessment, I looked to smartphone-based spectrophotometry as a potential new sensing avenue. I presented this poster on the topic at the UVA Engineering Research Expo and at the 2024 Commonwealth Cyber Initiative Symposium. Click the link for more details!

Constrained Reinforcement Learning

In 2023, I had the opportunity to work with PhD student Matt Landers at the HAI Lab on his project developing novel approaches to constrained reinforcement learning. Specifically, we aimed to find better ways of achieving knowledge transfer when the parameters of an environment may change. I helped implement several state of the art models, including policy distillation. The source code is available here.

Wirelessly-Powered Autonomous Warehouse Robots

For my junior research project at the Governor’s School, I designed and built a platform for warehouse robotics (like those used by Amazon) that were powered through wireless induction. Everything I did in this project, from programming and cloud infrastructure to computer-aided design and manufacture, was entirely self-taught. It placed first at the Tidewater Science and Engineering Fair and was accepted to the International Science Fair, but this was canceled due to COVID. Click the link for more details, or check out my writeup!

PeerCycle

I led a team of 4 on a semester-long project building PeerCycle: a platform for UVA students to list belongings that would normally be thrown away or left on the side of the road during move-out.

Gaggiuino

This last summer I took the time to perform the Gaggiuino mod on my espresso machine! Basically, this mod allows the machine to be controlled by a microcontroller and do some fancy stuff called ‘pressure profiling’ where it uses pressure, temperature, and weight sensors to pull the perfect shot. It was a lot of fun, but also a lot harder than I thought it would be (mostly because I thought I could swap out some recommended parts).

HAI Lab Website

I recently made a website for my lab group! You can check it out here to learn more about our work.