A  recent  area  of  interest  in  machine  learning  involves  drawing  inferences  from  a  large  number  of  agents,  each  
with some partial information.  Such problems arise in a number of settings, with recent areas of interest including 
recommender systems and sensor networks.  Problems in collective decision-making are also closely related to a 
fundamental problem of democracy — that of inferring the collective will of the people.  This talk will give a brief 
overview of machine learning and voting theory, followed by a discussion of some of our recent work in these areas.