BEGIN:VCALENDAR VERSION:2.0 PRODID:IEEE vTools.Events//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:DAYLIGHT DTSTART:20240310T030000 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:20240225T144050Z UID:1A27EFA8-3ACC-4CF1-9237-0EA5C6B8EE36 DTSTART;TZID=America/New_York:20240223T110000 DTEND;TZID=America/New_York:20240223T120000 DESCRIPTION:IEEE Communications Society Distinguished Lecture\, sponsored b y IEEE North Jersey ComSoc Chapter\, Systems Council and NJIT (host).\n\nS mart assistive environments adapt to the needs and preferences of disabled or elderly users who need help with the activities of daily living. Howev er\, the needs and requests of users vary greatly\, both due to personal p references and type of disability. As handcrafting an environment is prohi bitively expensive\, in recent years significant research was done in syst ems that use machine learning to create a predictive model of the user. Ma chine learning\, however\, typically requires large amounts of data. A sta nd-alone smart environment\, however\, only has access to the data collect ed from its user since it was deployed. A possible solution is to perform centralized\, cloud-based learning by pooling the training data collected from multiple users. However\, uploading data collected from the personal habits of elderly and disabled users create significant security and priva cy concerns.\n\nIn this talk\, we investigate the type of data sharing nec essary for learning user models in smart environments and propose several novel considerations. We point out that data sharing is only ethical if th e user derives a benefit from it. This implies that the decision to share data must be periodically revisited\, it is not a commitment extending ind efinitely in the future. We study the data sharing decisions made by users under several machine learning frameworks: local\, cloud\, and federated learning. We show that most users only benefit from data sharing for a lim ited interval after the deployment of the system. We also investigate mach ine learning techniques that predict whether the user will benefit from sh aring the data before the data is shared.\n\nCo-sponsored by: Systems Coun cil North Jersey\n\nSpeaker(s): \, Damla Turgut\n\nAgenda: \nSeminar from 11 am to noon. On-campus ECE 202 and zoom.\n\nRoom: 202\, Bldg: ECEC\, New Jersey Institute of Technology\, Newark\, New Jersey\, United States\, Vi rtual: https://events.vtools.ieee.org/m/405574 LOCATION:Room: 202\, Bldg: ECEC\, New Jersey Institute of Technology\, Newa rk\, New Jersey\, United States\, Virtual: https://events.vtools.ieee.org/ m/405574 ORGANIZER:a.j.patel@ieee.org SEQUENCE:15 SUMMARY:I Did Not Sign Up for This: Limited Sharing in Privacy-Aware Smart Environments URL;VALUE=URI:https://events.vtools.ieee.org/m/405574 X-ALT-DESC:Description: <br /><p>IEEE Communications Society Distinguished Lecture\, sponsored by IEEE North Jersey ComSoc Chapter\, Systems Council and NJIT (host).</p>\n<p>Smart assistive environments adapt to the needs a nd preferences of disabled or elderly users who need help with the activit ies of daily living. However\, the needs and requests of users vary greatl y\, both due to personal preferences and type of disability. As handcrafti ng an environment is prohibitively expensive\, in recent years significant research was done in systems that use machine learning to create a predic tive model of the user. Machine learning\, however\, typically requires la rge amounts of data. A stand-alone smart environment\, however\, only has access to the data collected from its user since it was deployed. A possib le solution is to perform centralized\, cloud-based learning by pooling th e training data collected from multiple users. However\, uploading data co llected from the personal habits of elderly and disabled users create sign ificant security and privacy concerns.</p>\n<p>In this talk\, we investiga te the type of data sharing necessary for learning user models in smart en vironments and propose several novel considerations. We point out that dat a sharing is only ethical if the user derives a benefit from it. This impl ies that the decision to share data must be periodically revisited\, it is not a commitment extending indefinitely in the future. We study the data sharing decisions made by users under several machine learning frameworks: local\, cloud\, and federated learning. We show that most users only bene fit from data sharing for a limited interval after the deployment of the s ystem. We also investigate machine learning techniques that predict whethe r the user will benefit from sharing the data before the data is shared.</ p><br /><br />Agenda: <br /><p>Seminar from 11 am to noon.&nbsp\; On-campu s ECE 202 and zoom.</p> END:VEVENT END:VCALENDAR