The word "personal" in personal computers should really attest to its purpose- to read the user's qualities and give corresponding feedback. Now, there seems to be devices that can literally read you. I read a paper published by the MIT media lab- "Capturing Individual and Group Behavior withWearable Sensors." The paper introduced a study that involved the sociometer and its reading of 67 nurses at a hospital in Boston. The sociometer recorded interactive information such as speech activity, face-to-face interaction, and physical activity. An accelerometer would read the individual's physical movements, and voice frequencies would recognize and analyze speech. Interactivity would be indicated by the interaction of two sociometers from 2 individual or by an individual's proximity to inanimate objects such as the phone or bed.
The results showed enough correlations to prove that sensor data can define personal traits and characteristics. It was interesting to see how people can be identified just based on physical qualities in relation other objects and humans. The data only had partial contextual references (based on one of the studies done in a setup) and barely incorporated the content of individual's thought process or conversations with others. This paper shows that perhaps the computer does not need intricate and overly complicated analysis of users to create an environment adapted for the user. More information must entail higher replica of the user's mind; however, raw sensor data could be sufficient to read human traits that are to some degree categorized.
Source:
"Capturing Individual and Group Behavior withWearable Sensors"
http://pubs.media.mit.edu/pubs/papers/AAAI2009_Human_Behavior_Modeling.PDF
No comments:
Post a Comment