HP Labs team wins first prize at international machine learning competition
A team of researchers from HP Labs and Michigan State University recently took home first prize in the prestigious “Discovery Challenge” of the 26th European Conference on Machine Learning and Principles and Practice of Knowledge Discovery, held in Riva del Garda, Italy.
The challenge asked entrants to make sense of data gleaned from four different types of sensors placed within a home environment and attached either to volunteers living in the home or placed around the building. The data was labeled by observers as representing 20 specific activities (walking, sitting, descending stairs, etc.). Some of the raw data was shared with the competitors matched to corresponding activity labels, while the rest was shared unlabeled. Each team was tasked with devising machine learning models and algorithms that could ‘learn’ from the labeled data to recognize the same activities in the unlabeled data in three activity groups: “ambulation”, “postures” and “transitions.”