Researchers at Carnegie Mellon University are taking on the challenge of building a continuously operating machine learning system that will attempt to master semantics by making use of the vast repository of natural language statements found on the Web. If successful, the system, known as NELL, for “Never-Ending Language Learner,” will result in a knowledge base of structured information that mirrors content on the Web. The team is using M45 – a supercomputing Hadoop cluster provided by Yahoo – to conduct their research. “The use of M45 made NELL a possibility that didn’t exist before,” said Tom Mitchell, head of machine learning at Carnegie Mellon. “We’re starting to look at even bigger scale computational issues involving iteratively reading billions of sentences. We’re thinking of M45 as a resource that makes it possible to think of such a crazy thing.” The project was featured in a recent New York Times article and a Tech Crunch video. You can also read more about the project here.