BEGIN:VCALENDAR
VERSION:2.0
PRODID:IEEE vTools.Events//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
DTSTART:20230312T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:EDT
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20231105T010000
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:EST
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20230919T204107Z
UID:9B130BE1-02A0-447D-A5CA-EE26552E652A
DTSTART;TZID=America/New_York:20230919T140000
DTEND;TZID=America/New_York:20230919T150000
DESCRIPTION:Smart assistive environments adapt to the needs and preferences
  of disabled or elderly users who need help with the activities of daily l
 iving. However\, the needs and requests of users vary greatly\, both due t
 o personal preferences and type of disability. As handcrafting an environm
 ent is prohibitively expensive\, in recent years significant research was 
 done in systems that use machine learning to create a predictive model of 
 the user. Machine learning\, however\, typically requires large amounts of
  data. A stand-alone smart environment\, however\, only has access to the 
 data collected from its user since it was deployed. A possible solution is
  to perform centralized\, cloud-based learning by pooling the training dat
 a collected from multiple users. However\, uploading data collected from t
 he personal habits of elderly and disabled users create significant securi
 ty and privacy concerns.\n\nIn this talk\, we investigate the type of data
  sharing necessary for learning user models in smart environments and prop
 ose several novel considerations. We point out that data sharing is only e
 thical if the user derives a benefit from it. This implies that the decisi
 on to share data must be periodically revisited\, it is not a commitment e
 xtending indefinitely in the future. We study the data sharing decisions m
 ade by users under several machine learning frameworks: local\, cloud\, an
 d federated learning. We show that most users only benefit from data shari
 ng for a limited interval after the deployment of the system. We also inve
 stigate machine learning techniques that predict whether the user will ben
 efit from sharing the data before the data is shared.\n\nSpeaker(s): Damla
  Turget\, \n\nRoom: 325B\, Bldg: ITE\, University of Maryland\, Baltimore 
 County\, 1000 Hilltop Circle\, Baltimore\, Maryland\, United States\, 2125
 0
LOCATION:Room: 325B\, Bldg: ITE\, University of Maryland\, Baltimore County
 \, 1000 Hilltop Circle\, Baltimore\, Maryland\, United States\, 21250
ORGANIZER:ashutosh.dutta@ieee.org
SEQUENCE:20
SUMMARY:Bringing privacy into the picture: new optimization goals for ML/AI
  in smart environments
URL;VALUE=URI:https://events.vtools.ieee.org/m/374300
X-ALT-DESC:Description: <br /><p>Smart assistive environments adapt to the 
 needs and preferences of disabled or elderly users who need help with the 
 activities of daily living. However\, the needs and requests of users vary
  greatly\, both due to personal preferences and type of disability. As han
 dcrafting an environment is prohibitively expensive\, in recent years sign
 ificant research was done in systems that use machine learning to create a
  predictive model of the user. Machine learning\, however\, typically requ
 ires large amounts of data. A stand-alone smart environment\, however\, on
 ly has access to the data collected from its user since it was deployed. A
  possible solution is to perform centralized\, cloud-based learning by poo
 ling the training data collected from multiple users. However\, uploading 
 data collected from the personal habits of elderly and disabled users crea
 te significant security and privacy concerns.</p>\n<p>In this talk\, we in
 vestigate the type of data sharing necessary for learning user models in s
 mart environments and propose several novel considerations. We point out t
 hat data sharing is only ethical if the user derives a benefit from it. Th
 is implies that the decision to share data must be periodically revisited\
 , it is not a commitment extending indefinitely in the future. We study th
 e data sharing decisions made by users under several machine learning fram
 eworks: local\, cloud\, and federated learning. We show that most users on
 ly benefit from data sharing for a limited interval after the deployment o
 f the system. We also investigate machine learning techniques that predict
  whether the user will benefit from sharing the data before the data is sh
 ared.</p>
END:VEVENT
END:VCALENDAR