Machine Learning and Friends Lunch





home
past talks
resources
conferences

Learning Hierarchical models of activity


Sarah Osentoski
UMass

Abstract


In this talk I will discuss a method of learning hiearchical statistical activity models in indoor environments. The Abstract Hidden Markov Model is used to represent behaviors in these stochastic environments. We train the model using both labeled and unlabeled data and estimate the parameters using EM. Results are shown on three datasets: data collected in lab, entryway, and home environments.

Back to ML Lunch home