Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Health Informatics Thursday, November 29, 2012: Active Sensing for Partially Observable Markov Decision Processes
1 view
Skip to first unread message
wlr...@uwaterloo.ca
Nov 23, 2012, 2:23:21 PM11/23/12
to
A new Health Informatics event has been scheduled:
Active Sensing for Partially Observable Markov Decision Processes
Context information on a smart phone can be used to tailor applications
for specific situations (e.g. provide tailored routing advice based on
location, gas prices and traffic). However, typical context- aware smart
phone applications use very limited context information such as user
identity, location and time. In the future, smart phones will need to
decide from a wide range of sensors to gather information from in order
to best accommodate user needs and preferences in a given context. In
this thesis, we present a model for active sensor selection within
decision-making processes, in which observational features are selected
based on longer-term impact on the decisions made by the smart phone.
This problem is more challenging than measuring information gain,
because context information gathering can improve the quality of the
decisions made at some point in the future. This thesis formulates the
problem as a partially observable Markov decision process (POMDP), and
shows how the POMDP can integrate partially available sensor information
with user goals and preferences, allowing a seamless fusion of uncertain
sensor data with complex and long-term decision-making on the mobile device.
Date: Thursday, November 29, 2012
Time: 9:00
Please look on WebCalendar to view this appointment.
http://www.cs.uwaterloo.ca/odyssey/event/1631
Active Sensing for Partially Observable Markov Decision Processes
Context information on a smart phone can be used to tailor applications
for specific situations (e.g. provide tailored routing advice based on
location, gas prices and traffic). However, typical context- aware smart
phone applications use very limited context information such as user
identity, location and time. In the future, smart phones will need to
decide from a wide range of sensors to gather information from in order
to best accommodate user needs and preferences in a given context. In
this thesis, we present a model for active sensor selection within
decision-making processes, in which observational features are selected
based on longer-term impact on the decisions made by the smart phone.
This problem is more challenging than measuring information gain,
because context information gathering can improve the quality of the
decisions made at some point in the future. This thesis formulates the
problem as a partially observable Markov decision process (POMDP), and
shows how the POMDP can integrate partially available sensor information
with user goals and preferences, allowing a seamless fusion of uncertain
sensor data with complex and long-term decision-making on the mobile device.
Date: Thursday, November 29, 2012
Time: 9:00
Please look on WebCalendar to view this appointment.
http://www.cs.uwaterloo.ca/odyssey/event/1631
0 new messages