Ready to welcome your new robot overlords?
Fear not — we’re far from Terminators taking over. But artificial intelligence is indeed entering our daily lives, with many useful benefits that might surprise you. Here are five ways USC scientists put machines to work for all humankind.
USC researchers devised a new way to fight back against the novel coronavirus by pinpointing the most promising vaccines and therapies in seconds — keeping us ahead of its mutations. Their machine-learning approach analyzes potential changes in the virus that causes COVID-19 and zeros in on the best vaccine candidates.
Ever wonder how a movie gets its PG-13 rating? Usually, it falls to human reviewers to decide after the movie is filmed. But USC scientists built an artificial intelligence tool that can pick out mature material in film scripts before any scenes are shot. That can help storytellers reflect on the messages they want to send and whether they are limiting their potential audience.
Picture of Health
USC researchers are turning to artificial intelligence methods to comb through massive databases of information from Alzheimer’s patients to identify early signs and potential causes of the disease. Scientists are also building a machine-learning tool to help radiologists analyze kidney scans for potential tumors, enabling physicians to identify and treat renal cancer earlier than ever.
Alexa and Google are great at understanding commands — playing a song or turning down the lights, for example — but their conversation skills could use some work. USC experts are hoping artificial intelligence can learn from an unexpected source of dialogue: improv comedy. By feeding thousands of comedic exchanges into a chatbot system, the researchers hope to create a fun and witty conversation partner that you’d never guess is powered by AI.
As humans, dreaming up things we’ve never seen — say, a purple cat with orange eyes — starts in childhood. Ask a computer to use its imagination, however, and it might hit a creative roadblock. But USC scientists are on the right track: They recently created a new form of AI that can extrapolate from the knowledge it’s taught. The researchers hope that this could make for safer artificial intelligence, like helping self-driving cars imagine and dodge dangerous situations they’ve never encountered before.