Brazilian EdTech Startup Collaboration

Nicolas Perez *Wesley Tong,  Neil McWhorter,  Isaac Gerstmann,  Morgan Lu,  Mengfei Zhang,  Matthew Cho
* Project Lead

A CAIS++ team worked with AIO, a Brazilian education technology startup. AIO develops low-cost, personalized, AI-powered test preparation to help students study for the Exame Nacional do Ensino Médio (ENEM), a national standardized Brazilian exam that determines university admission and scholarship opportunities for high school students in Brazil. Our CAIS++ team worked on deep, transformer-based knowledge tracing models that contributed to AIO's platform.

Viral Genomics for Predicting Pandemics (2.0)

Shannon Brownlee *,  Sam Sommerer *Robb Tran,  Elaine Yang,  Anthony Martino,  Jordan Cahoon
* Project Lead

This research project at the intersection of ML and genomics focuses on analyzing virus genome sequences, mutations, and behaviors to predict the pandemic risk potential of a virus. This project was the second phase of the Fall 2020 CAIS++ viral genomics project.

Clinical Alzheimer's Dementia Diagnosis

Leena Mathur *,  Nisha Chatwani *Surya Nehra,  Maxwell Kofman,  Jessie Zhang,  Nicolas van Houten
* Project Lead

This research developed machine learning models that leveraged psycholinguistic and acoustic speech representations for low-cost, scalable Alzheimer's dementia diagnosis. This group led the first cross-USC research collaboration between CAIS++ and another student organization (MEDesign biomedical engineers). This CAIS++ group also presented at USC's Spring 2021 undergraduate research symposium and won the Interdisciplinary Prize, sponsored by the Office of the Provost/USC Fellowships Office.

Detecting Wildfires

Ishu Agrawal *Lauren Tsai,  Pau Sang,  Lauren Okhovat
* Project Lead

This project developed automated and scalable computer vision models for wildfire detection that can be used with data streams from aerial photography or forest monitoring stations.

Embedded Computer Vision

Shantanu Jhaveri *,  John Bush *Allen Chang,  Arman Roshannai,  Avi Gala,  Megan Friedenberg,  Pilar Luiz,  Samuel Mesfin,  Wesley Tong
* Project Lead

This CAIS++ team experimented with the Oak-D, an smart stereo camera capable of running complex neural network models for computer vision tasks. This CAIS++ group learned about state-of-the-art object recognition and detection models, and they developed a system for real-time perception of whether or not people were wearing masks in social spaces.

FALL 2020

ProjectX: Inefficiencies in Renewable Energy

Patrick Darrow *,  Nicolas Perez *Tomas Angelini,  Christopher Fucci,  Priscilla Lee,  Gireesh Mahajan
* Project Lead

A CAIS++ team represented USC at ProjectX, a machine learning research competition hosted by the University of Toronto. This competition was focused on tackling climate change problems with ML, and our group addressed inefficiencies associated with variable renewable energy sources.

Viral Genomics for Predicting Pandemics (1.0)

Shannon Brownlee *,  Laura Cao *Gloria Chang,  Shantanu Jhaveri,  Karan Menon
* Project Lead

This research project at the intersection of ML and genomics focuses on analyzing virus genome sequences, mutations, and behaviors to predict the pandemic risk potential of a virus.

Debiasing ML Recruiting Models

Ritika Dendi *,  Nico van Houten *Brittany Rollins,  Mingtao Dong,  Natalie Abreu,  Samuel Sommerer
* Project Lead

Existing ML-based recruiting tools have been biased against people from historically-underrepresented backgrounds. This research tackles the problem of biased ML recruiting models in the context of resume evaluations by exploring and implementing solutions such as hard debiasing and adversarial debiasing.

Addressing K-12 School Performance Inequalities

Leena Mathur *Andrew Hariri,  Nicole Ng,  Jae Shim
* Project Lead

Achievement gaps persist among K-12 students and schools across the United States, often correlated with population-level socioeconomic, demographic, and geographic characteristics of a school district, as well as school-level factors such as school funding amounts and student to teacher ratios. This research leverages ML to identify key features predictive of school performance, in order to develop data-driven community interventions at the school district level and school level to boost school academic performance. This research was advised by Professor Bistra Dilkina, and findings are currently under preparation for publication.

The Ladin Language: Preserving Endangered Languages through Phoneme Recognition (4.0)

Zane Durante *,  Leena Mathur *Eric Ye,  Jack Zhao,  Tejas Ramdas
* Project Lead

Approximately ~60% of the world’s current 7,000 languages is predicted to go extinct by the end of the century; the death of any language represents an irreversible loss of information across multiple fields, including linguistics, psychology, sociology, and anthropology. A CAIS++ team continued their work on developing unsupervised representation learning techniques for automatically extracting and preserving the phonemes of endangered languages, through a collaboration with USC's Department of Linguistics. This team also wrote a paper from their research findings!