What if an algorithm existed that breaks down your energy bill appliance by appliance?
That's the idea behind the research of Zico Kolter. Kolter believes that if the average consumer understood exactly how much energy their fridge, washer and dryer and lights added to their overall energy consumption, they would work harder at conserving energy.
A native of Boston, Kolter attended Georgetown (undergrad) and Stanford (grad). As an assistant professor in CMU’s Computer Science Dept. and Institute for Software Research, he is offering a new course this fall that will explore the ways machine-learning can address this idea.
Kolter says machine-learning is the answer to a clearer understanding of energy consumption. Energy is the driving force of human society. If people could actually see where it goes everyday, they could potentially reduce their consumption by 15%. And that's a conservative estimate.
Businesses and industries could achieve a 25% reduction, he says.
His research team—civil, electrical and chemical engineers—are busy at work on new computational approaches. Kolter hopes to collaborate with a large energy company in the near future, as well, and put advanced consumer energy technology to the test.
“Source separation is hard to tackle, but can be done,” he says. “The big take home is we want to give people information about their consumption. It costs money to develop new sources of technology. It’s relatively cheaper to be more efficient with the energy we are generating.”
The right algorithms can work to support alternative forms of energy like wind and solar, he adds.
Writer: Deb Smit
Source: Zico Kolter, CMU