human‐machine coexistence in groups
DESCRIPTION
Smart Society George Kampis & Stuart Anderson: Edinburgh University, Paul Lukowicz: DFK Presentation from ECAL 2013TRANSCRIPT
Human‐machine Coexistence in GroupsGroups
George Kampis Paul Lukowicz Stuart Anderson DFKI (German Research Institute for Artificial Intelligence)DFKI (German Research Institute for Artificial Intelligence),
Kaiserslautern, Germany
Edinburgh University, Edinburgh, United Kingdom
26/09/2013 www.smart‐society‐project.eu
“ l l” l“classical” telemonitoring
hScoring scheme
lResults Patients like the system – Scoring is de‐personalisedPatients like the system primarily because it gave them privileged access to doctors...
Doctors orried abo t
Scoring is de personalised Contextual factors are excluded
Doctors worried about: High false positive rate Over diagnosis
Difficult to prioritize patients Brittle response even at small scale
Over treatment
“In participants with a history of admission for exacerbations
small scale
of admission for exacerbations of COPD, Telemonitoring was not effective in postponing hospital admissions orhospital admissions or reducing healthcare costs.”
Diagnosis Capturing Context Shifting Context Capturing Context Pervasive and mobile data capture
Shifting Context Individual Global systemicdata capture
Formal monitoring activity
Global, systemic Individualised
Open‐ended Social Negotiation of key features
26/09/2013 www.smart‐society‐project.eu 5
Diagnosis Cultures Learning Cultures Domestic: concerned, inexpert
Learning More or less open loop Easy to createinexpert
Call Centre: risk averse, inexpert
Easy to create “revolving door” cases.
General Practitioner: risk averse, expert, resource constrained
Acute Care: uninvolved
26/09/2013 www.smart‐society‐project.eu 6
“ l k ”“Social Sensemaking” Individual family and carers “curate” the monitoring time Individual, family and carers curate the monitoring time series.
Capture rich context – environment patient condition Capture rich context – environment, patient condition, … Learning algorithm Interactive supports negotiation Interactive, supports negotiation Linking cultures supporting the transfer of context Individualised Individualised
Learning communityH i t l li k Horizontal links
Transferring experience
26/09/2013 www.smart‐society‐project.eu 7
Settings Cultures Context Cultures Long‐lived Value rich (empirical
Context Rich Open ended Value rich (empirical
work) Range of scale
Open ended Unexpected Capture is inherentlyg
Shifting and overlapping Capture is inherently hybrid
Sensors always miss ythings
Negotiated
26/09/2013 www.smart‐society‐project.eu 8
hMechanisms Semantics Mediation Semantics Tightly linked to culture/context
Mediation Value systems differ Boundary practicesculture/context
“Good enough” to understand now
Boundary practices enable inter‐working without agreement
Open‐ended Structured by cultures
Key element in crossing cultures
26/09/2013 www.smart‐society‐project.eu 9
hMechanisms Incentives Hybridity Incentives Easy to get wrong E g quality of care in
Hybridity All computation is social In particular the E.g. quality of care in
the telemonitoring case “Keeping people out of
In particular the evolution process
Develop more p g p phospital” is probably better
pparticipatory approaches to
Heterogeneous Privacy harming?
evolution.
26/09/2013 www.smart‐society‐project.eu 10
ffEffects Learning Alignment Learning Opportunity to reflect Incorporate reflection
Alignment Temporary, requires repeated repair Incorporate reflection
Drives design and evolution
repeated repair Incentives encourage alignment
Linkage between values and incentives
26/09/2013 www.smart‐society‐project.eu 11
lConclusions Focus on Hybridity and Diversity Focus on Hybridity and Diversity Hybridity drives the programming model…Di it d i th d t d l Diversity drives the data model…
26/09/2013 www.smart‐society‐project.eu 12