Round-Up: Recent USC Viterbi Research
Cooling Your Home Earlier Is Cool for the Environment
Professor Kelly Sanders and Ph.D. student Stepp Mayes found that dramatically cooling homes with renewable energy in the middle of the day, before demand surges around 4 p.m., could reduce carbon dioxide emissions, save consumers money and prevent rolling blackouts. Specifically, pre-cooling could cut a typical home’s air conditioning energy consumption during peak periods by roughly 40% to 50%, reduce carbon dioxide emissions and save customers up to $30 per month.
Putting Cancer Under a Bright Light
A new fluorescent probe developed by Professor Andrea Armani and Ph.D. student Yasaman Moradi illuminates protein clusters found in a subset of breast cancers. Their new tool both detects the presence of and measures the spatial distribution of HER2 on the cell, an improvement over existing methods that could hasten the development of new drugs to treat breast cancer.
Pro-Tumor or Anti-Tumor?
Professor Stacey Finley and Ph.D. candidate Patrick Gelbach studied immune cells, known as macrophages, surrounding colorectal tumors. These cells can have anti-cancer properties, or they can be “pro-tumor,” aiding cancer proliferation. The work identified unique metabolic reactions that can be harnessed to convert the cells from pro-tumor to anti-tumor, to block the cancer’s spread.
The Future of Farming
Associate Professor Barath Raghavan is rethinking industrial farming practices by developing computational tools to help farmers design, develop and manage sustainable farming methods. From crop selection to planting to irrigation, his new computational method allows farmers to explore thousands of different potential designs to optimize food production and create sustainable agricultural ecosystems.
AI-powered Sign Language Detection
Sign language recognition and translation technologies have the potential to increase access and inclusion for deaf signing communities, but research is bottlenecked by a lack of representative data. Lee Kezar, a computer science doctoral student advised by Assistant Professor Jesse Thomason, created a dataset with 84,000 videos of isolated sign productions from American Sign Language users. This could be an invaluable resource for the advancement of AI-powered sign language recognition technologies.