‘Liquid’ machine-learning system adapts to changing conditions

 Daniel Ackerman of MIT in TechXplore –

neural network

MIT researchers have developed a type of neural network that learns on the job, not just during its training phase. These flexible algorithms, dubbed “liquid” networks, change their underlying equations to continuously adapt to new data inputs.

The advance could aid decision making based on data streams that change over time, including those involved in medical diagnosis and autonomous driving.

“This is a way forward for the future of robot control, natural language processing, video processing—any form of time series data processing,” says Ramin Hasani, the study’s lead author. “The potential is really significant.”

The research will be presented at February’s AAAI Conference on Artificial Intelligence.

Continue…

About Radnor Reports

Ken Feltman is past-president of the International Association of Political Consultants and the American League of Lobbyists. He is retired chairman of Radnor Inc., an international political consulting and government relations firm in Washington, D.C. Known as a coalition builder, he has participated in election campaigns and legislative efforts in the United States and several other countries.
This entry was posted in Controversial. Bookmark the permalink.