Keenan Crane Describes Differences in Contemporary ML Models

Keenan Crane, an associate professor of computer science and robotics at CMU, posted a reply outlining two major differences in contemporary machine learning models. He stated that distributions in these models are learned from data rather than defined by hand. He added that the models are far more general because they incorporate cross-domain knowledge instead of being tailored to one domain. Crane concluded that these changes amount to a big deal. The post appears as a single reply in an ongoing conversation about artificial intelligence on the platform.