Date post: | 05-Dec-2014 |
Category: |
Technology |
Upload: | focas-initiative |
View: | 219 times |
Download: | 2 times |
. . .for Socio-Technical Systems
Jeremy PittDepartment of Electrical and Electronic Engineering
Imperial College London
FoCAS Summer SchoolUniversity of Crete, Crete, 23-27/06/2014
Jeremy Pitt . . . for Socio-Technical Systems 1 / 34
Framing the Problem (1)
Dear all,
Faculty members have received complaints by several peopleof very noisy and other unpleasant (smelly due to food)conditions in some of the working spaces (PhD and RArooms). In general, we do not like micromanaging and we letpeople resolve their small problems and annoyances amongthemselves through reasonable discussions that we assumethey are all capable of.
(Names removed to protect the innocent as well as the guilty)
Jeremy Pitt . . . for Socio-Technical Systems 2 / 34
Framing the Problem (2)
Jeremy Pitt . . . for Socio-Technical Systems 3 / 34
Framing the Problem (2)
Jeremy Pitt . . . for Socio-Technical Systems 4 / 34
Framing the Solution
Unpicking the problem
“complaints” ⇒ an adverse affective response and diminished‘quality of experience’ (QoE)“do not like micromanaging” ⇒ events at a low level ofsignificance except for those impacted“among themselves” ⇒ restoration and maintenance of ahomeostatic equilibrium based on decentralised collectivedecisionsembedded in the context of social interaction ⇒ representationand reasoning about norm-governed behaviour fordecision-support (i.e. social intelligence)
Collective Action problem
Actions which provision to and/or appropriate from a commonpoolShare a common resource with competing and conflictinginterestsActions with individual costs but group benefits
Jeremy Pitt . . . for Socio-Technical Systems 5 / 34
Setting the Goal
To create a generic socially-intelligent socio-technical systemfor enhanced QoE in shared physical spacesusing collective adaptive systemsover time, and space
Algorithmic self-governance: the structures, functions andprocesses by which an intelligent system embedded in ahuman techno-social environment can configure (andreconfigure) itself for improved QoE or other aspects ofqualitative human concern
Jeremy Pitt . . . for Socio-Technical Systems 6 / 34
Agenda
Motivation
Framing the problemFraming the solution
Background
PsychologySocial ScienceComputing
Experiments (Feasibility Studies)
Emotive EpisodeEnhanced QoE in Public CollectionsCuing Self- and Social-AwarenessSensor Networks for Offices
Socially-Intelligent Workplace System: Design and Evaluation
Summary and Conclusions
Jeremy Pitt . . . for Socio-Technical Systems 7 / 34
Background
Jeremy Pitt . . . for Socio-Technical Systems 8 / 34
Background – Psychology
The design of workplaces has a profound influence onwork-related issues, such as productivity and efficiency
Even an ideal physical arrangement of machines andworkstations is subject to:
the social and emotional intelligence of the people using thempersonnel churn, fast-paced turn-over in office occupancy, andmediation through high-tech, asynchronous communications
The result (potentially) is uncivil behaviour:
“low-intensity deviant behaviour with ambiguous intent toharm the target, in violation of workplace norms for mutualrespect. Uncivil behaviours are characteristically rude anddiscourteous, displaying a lack of regard for others.”
The prevailing and costly effects of workplace deviance areconsidered to be one of the most serious problems thatorganisations currently face
Jeremy Pitt . . . for Socio-Technical Systems 9 / 34
Affect and Affective Computing
Emotions play a critical role in cognitive processes in humans
Workplace interactions might be improved by leveraging theemotions of each of the individuals that are part of theworkplace, with technologyPsychophysiology
Classification of affective stateEkman and Friesman cross-culural study of facial recognitionidentified 6 basic emotionsPlutchik’s conic model with primary and secondary affectivestatesRussel’s emotional space, including positive and negativevalues of valence and arousal
Affective computingphysiological signals are an indicator of psychological (emotive)statesignal processing techniques can be used, in conjunction with atheory of emotion to infer that emotional state from recordedsignals
Jeremy Pitt . . . for Socio-Technical Systems 10 / 34
Illustration
Clockwise, from top
anger, fear, disgust,surprise, happines andsadnessPlutchik’s Wheel ofEmotionsRussel’s circumplexmodel
Jeremy Pitt . . . for Socio-Technical Systems 11 / 34
Forgiveness
Interaction between anonymous parties in one-off encounterssimplistic game-theoretic analysis suggests optimal strategy isdefectin iterated encounters, once defect is ‘played’, there is noredemptionyet this is not always the behaviour we observe
Encourage compliance by provision of cues, but:norm violation is a trust breakdownreputation management is a punishment mechanism, not areparation mechanismdoes not distinguish between intentional and unintentionalviolations
Inability to escape mutual ‘tit-for-tat’ exacerbated by lack ofcues
low self-awareness in certain settings hinders an offender’sexperience of shame, guilt or embarrassment;want instead to activate self-conscious emotions which in turnmotivate the cessation and reversal of offensive behaviour
Jeremy Pitt . . . for Socio-Technical Systems 12 / 34
Background (2) – Social science
From manager’s perspectiveFor an ‘office procedure’ to achieve its objectives it isnecessary to understand the impact on policy subject’s(subjects’) behaviourParticipatory adaptation for management of common poolresources (CPR)
From occupant’s perspectiveParticipatory adaptation for management of common poolresources (CPR)
People do not comply/not-comply with regulations, they reactto incentives implied by the regulations
Agreed rules and well-defined procedures, and rules andprocedures for changing rules and proceduresWays of referring to, reasoning about, and interacting withothers concerning those rules and proceduresFormalise through roles and social constraints (power,permission, obligation)
Jeremy Pitt . . . for Socio-Technical Systems 13 / 34
Experiments (Feasibility Studies)
Jeremy Pitt . . . for Socio-Technical Systems 14 / 34
The Emotive Episode (with Farrer and Goulev)
Application of the AffectiveWare SystemIncluding devices for measuring physiological signals: rings andmouse
The experiment itself: two parts:The users held 4 pieces of iconic branded fashion garmentswhilst reading the company PR from the official web site. Thegalvanic skin response was measuredThe experiment part two. The users were shown the samepieces with a brief explanation of a sustainable issue in thepublic domain connected to each item relative to clothingsupply chains consumption and disposal. The galvanic skinresponse was measured
ResultsThere was a measurable physiological response to holdingiconic clothing itemsThe response was different if the subjects were readingcompany PR as opposed to factual information aboutsustainable issues while holding the item
If we cannot infer actual affective state, we can infer changeof stateJeremy Pitt . . . for Socio-Technical Systems 15 / 34
AffectiveWare
The affective rings which measure galvanic skin response usedin the experiment
Jeremy Pitt . . . for Socio-Technical Systems 16 / 34
QoE in Public Collections (with A. Bhusate)
Problem: public collections are under pressure to maintainvisitor numbers
Solution: enhance the Quality of Experience (QoE) in a visitto a museum
Leverage pervasive computing with visitors’ own devices
saturate (pervade) the environment, including the exhibits,with sensors and displaysdefine a policy language to allow customisation according tothe user (visitor)use ‘intelligent’ decision-making to adapt behaviour/displaybased on behaviour/data fusion with respect to the policystream material direct to the visitor’s device
Jeremy Pitt . . . for Socio-Technical Systems 17 / 34
Interactive Exhibit
Usability lab, with interactive exhibit loaded with sensors:touch sensors, micro-switches, compass, 3-axis x-y-zaccelerometer, bluetooth connection
Driver Door Open/ Close Sensor
Engine Touch Sensor
Boot Sensor
Driver Seat Touch Sensor
Honeywell HMC6352 Compass PIC16f690 Micro - processor
LIS3LV02DQ 3-Axis Accelerometer T9JRN41 Bluetooth Module
Gear Touch Sensor
Steering Wheel Touch Sensor
Bonnet Open/ Close Sensor
Jeremy Pitt . . . for Socio-Technical Systems 18 / 34
Experiments with Affect and Forgiveness in CMC (with A.Vasalou)
Experiment 1: Interface Cues to simulate self-awarenessUse interface mechanisms to activate self-conscious emotions,which in turn motivate the reversal of offensive behaviour
Avatars as one mechanismFor example, private self-awareness can be manipulated withevaluative cuesConcept built into an avatar which autonomously expressesvisible embarrassment/shame upon its owner’s offenceWhen expressing these emotions, the avatar changes in a stylewhich is reported as common for embarrassment and shame
Jeremy Pitt . . . for Socio-Technical Systems 19 / 34
Forgiveness Framework
Experiment 2: forgiveness intervention mechanism forrepairing unintentional violations
Decision Maker based on Fuzzy Inferenceeleven constituent signals underlying the four positivemotivationsa formula used to compute strength of each signalinput into fuzzy inference rules defining FIS2..FIS5
if severity is low and frequency is low and intent is highthen judgment of offence motivation is 0.4
Combined into a final decision by FIS1
Jeremy Pitt . . . for Socio-Technical Systems 20 / 34
Office Sensor Networks
Filming a workplace
Situations that upsetthe work stream.How people react tothose situations.
Cameras were placed in aworkplace and filmed itduring a work week
1 camera to film thewhole room;8 cameras to filmindividuals
Analysis of recordings +interviews of workplaceelements
Establishment of socialnorms for the workplace.
Jeremy Pitt . . . for Socio-Technical Systems 21 / 34
Observations from Feasibility Studies
Emotive episode suggests that we can detect change of useraffective state
Sensor networks can be used to detect user actions and relatethem to norms specified in a policy
Interface mechanisms can be used to promote self-awarenessfor compliant behaviour and reparations of unintentionalviolation
Offices can be electronically-saturated environments
Bring all this together to build a system to enhance QoE inworkplaces
Jeremy Pitt . . . for Socio-Technical Systems 22 / 34
Socially-Intelligent Workplace System
Jeremy Pitt . . . for Socio-Technical Systems 23 / 34
Open Plan Office Questionnaire
Target respondents: people who share or have shared aworkplaceResponse demographics
125 respondents, 37 females, 88 malesmostly from (85.6%) and working in (97.6%) Europeaged between 22 and 47 years old
Objective: to understandwhich behaviours are considered to be uncivil and harmful tothe good environment of the workplaceself-awareness (which self behaviours might be uncivil)personal techniques for dealing with others’ uncivil behavioursexisting social norms
Preliminary results:
noise seems to be the mostprevalent issueother uncivil behaviours: smellyfood, affecting environmentwithout asking, . . .
Jeremy Pitt . . . for Socio-Technical Systems 24 / 34
MACS (M—’s Affective Conditioning System)
Provides a computer-mediated interaction between people in aworkplace
The defined social norms are mapped in the system through apolicy-based language
The violation of norms is reported by individuals anonymously,to avoid lack of participation caused by inhibition
There are decision-making tools to control the providedfeedback
Offenders have the opportunity to apologise and/or explaintheir actions
Analogy is with air conditioning system - restoration of ahomeostatic equilibrium
Jeremy Pitt . . . for Socio-Technical Systems 25 / 34
MACS Architecture
Jeremy Pitt . . . for Socio-Technical Systems 26 / 34
MACS Interface Functionality
Each computer in the workplace has a client interface of thesystem
Each person can use their instance of the system to:
Read the social norms of the workplaceCheck their own behaviour historyFlag a violation of social norms or an action that upsets thework streamWrite an explanation/apology for breaking the social norms (ifs/he broke the norms)Check redeeming inputs from offendersChange ‘opinion’ about offenders’ actions that were flagged bythem
Jeremy Pitt . . . for Socio-Technical Systems 27 / 34
MACS Interface – Social Norms
Those affected by provision and appropriation rules shouldparticipate in their selection
Jeremy Pitt . . . for Socio-Technical Systems 28 / 34
MACS Interface – Personal View
Worker’s view
Jeremy Pitt . . . for Socio-Technical Systems 29 / 34
MACS ‘Back Office’ Functionality
The main system is installed in a separate computer
The administrator interface shows a map of the workplace,with each person being represented by a chosen avatar
Provides an overview of all the events
What violations are flagged as offencesWho are the offendersWho flagged each offenceWhat feedback is provided to whom
Jeremy Pitt . . . for Socio-Technical Systems 30 / 34
MACS Interface – Administrator View
Administrator Interface
Jeremy Pitt . . . for Socio-Technical Systems 31 / 34
Evaluation Outcomes
Social connectedness evaluation
Evaluate feeling of belonging before and after the use of MACSQuestionnaire asking how well people know each other andhow much they feel they belong to the workplace
Quality of Experience evaluation
Open interviews about the effects of the introduction of theMACS in the workplaceStudy of workplace environment indicators (in MACS)
Evolution of number of reported offencesEvolution of the reputation of each person of the workplace
Aim to have something concrete on guidelines for use of ICTin workplace design to improve ‘sense of belonging’,participatory adaptation and enhanced QoE
Jeremy Pitt . . . for Socio-Technical Systems 32 / 34
Summary
Addressing a substantial problem in today’s knowledgeeconomy given specific social, cultural and organizationalpressures
Perspective: shared environment as a common-pool resource
Attempt to create socially-intelligent ICT convergingnormative+affective+pervasive+adaptive computing
Dispute resolution and conflict prevention using autonomicsocio-technical systems
Can we improve QoE and social-connectedness in workplaceenvironments? . . .
What do formal models of ‘empathy’ look like?
Visualisation of ‘moral’ design choices Longrightarrow designcontractualism
Can we build self-governing institutions for socio-technicalsystems? . . . (cf. ID3 and the Open Mustard Seed (OMS)project)
Jeremy Pitt . . . for Socio-Technical Systems 33 / 34
Bibliography
L. Andersson and C. Pearson, Tit for tat The spiraling effect of incivility in the workplace, Acad. Manag.Rev., pp. 452-471, 1999.
R. Bennett and S. Robinson, The past, present, and future of workplace deviance research, Org. Behav.:State Sci. 2, pp. 247-281, 2003.
P. Ekman and W. Friesen. Unmasking the Face: A Guide to Recognizing Emotions from Facial Clues.Prentice-Hall, 1975
J. Farrer, P. Goulev, J. Pitt. The Emotive Episode: an investigation into user response to SustainableIssues in Fashion/Textiles and Affective Computing, Sustainable Innovation 06, Chicago, 2006
C. Hewitt. Offices Are Open Systems. ACM Trans. Inf. Syst. 4(3): 271- 287, 1986
E. Ostrom. Governing the Commons: the Evolution of Institutions for Collective Action. CUP, 1990
C. Pearson and C. Porath, On the nature, consequences and remedies of workplace incivility: no time fornice Think again, Acad. Manag. Exec. 19 (1), pp. 7-18, 2005.
R. Plutchik. Emotion: A psychoevolutionary synthesis. Harper & Row New York, 1980
J. Russell. A Circumplex Model of Affect. Journal of Personality and Social Psychology, 39(6):11611178,1980
A. Seidman and R. Seidman, ILTAM: Drafting Evidence-based Legislation for Democratic Social Change.Boston University Law Review 435, 2009.
A. Vasalou, A. Joinson, J. Pitt. Constructing my online self: avatars that increase self-focused attention.In proceedings of CHI 2007. CHI Notes. San Jose, USA, 2007
A. Vasalou, J. Pitt. Reinventing forgiveness: a formal investigation of moral facilitation. In proceedings ofthe iTrust Conference. Paris, France, 2005.
Jeremy Pitt . . . for Socio-Technical Systems 34 / 34