Activity Recognition

One of my beefs with activity recognition in the machine learning world is a lack of understanding of how it will be used. It is often suggested that it would be useful in a variety of vague ways, elder-care, security, etc., but to really do activity recognition you have to get into the details. I don't think that machine learning researchers really like to get into the details of an application. It is generally seen as a weakness because it isn't sufficiently general to transfer to new domains.

One of my students, Sharon Ding, researched this great quote that solidifies some of my hesitations with general activity recognition. It is by an American philosopher, George Mead who was writing in 1934:

"the behavior of an individual can be understood only in terms of the behavior of the whole social group of which he is a member, since his individual acts are involved in larger, social acts which go beyond himself and which implicate the other members of that group."