Machine learning is now in everything from smartphones that listen to voice commands to cars that drive themselves. But algorithms still fall far short of humans’ abilities to learn—the algorithms require huge amounts of training data and often make decisions that defy common sense.
IU researchers are investigating machine learning from many perspectives, including studying its theoretical properties and limitations; developing new algorithms and models; improving scalability for large, noisy data; understanding the connections to human learning; and applying machine learning to a wide variety of problems.