With emerging technologies and big data, it is now possible to design intelligent social systems. In this presentation, ideas related to designing such systems are presented
2. • Social
systems
rely
on
primi0ve
technology.
• Big
Data
has
opened
Big
Opportuni0es.
• Situa0on
recogni0on
is
a
key
technology.
• EventShop
may
be
useful
in
designing
Intelligent
Social
Systems.
4. Intelligent:
displaying
or
characterized
by
quickness
of
understanding,
sound
thought,
or
good
judgment.
Social
Systems:
Social
systems
are
the
paBerns
of
behavior
of
a
group
of
people
possessing
similar
characteris1cs
due
to
their
existence
in
same
society.
5. • Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Social
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
6. An
Interes0ng
Situa0on
When
we
were
data
poor
–
we
searched
for
words
in
documents.
Now
that
we
are
data
rich
–
should
we
s0ll
search
for
words?
Time
has
come
for
us
to
stop
thinking
data
poor;
really
start
thinking
and
behaving
data
rich.
7. Variety
Volume
Big
Data
offers
Big
Opportuni4es.
But,
….
?????
7
8. Most
aOen0on
by
Top
1.5
Technologists
–
so
far.
Billion
Middle
4
Billion
Middle
of
the
Pyramid
(MOP):
Ready.
BoOom
2
Billion
Not
Ready
9. Data
is
Essen0al.
But,
we
are
really
interested
in
its
products:
Informa0on,
Knowledge,
and
Wisdom.
9
10. Recognize
Objects
Situa0ons
Knowledge
Observe
Big
Data
Act
Planning
12/5/12
Control
10
11. Past is EXPERIENCE
Present is EXPERIMENT
Future is EXPECTATION
Use your Experiences
In your Experiments
To achieve your Expectations
12/5/12
11
12.
Astrology
To
Astronomical
Volumes
of
Data
12/5/12
12
13. We
are
immersed
in
Networks
of
• People
• Things
• Events
It
is
now
possible
to
be
Pansophical.
12/5/12
13
14. Our
mobile
wireless
infrastructure
can
be
“reality
mined”
to
understand
the
paOerns
of
human
behavior,
monitor
our
environments,
and
plan
social
development.
-‐-‐-‐-‐
Pentland
in
“Society’s
Nervous
System:
Building
Effec0ve
Government,
Energy,
and
Public
Health
Systems”
Proprietary
and
Confiden1al,
Not
For
12/5/12
14
Distribu1on
15. • Objects
-‐-‐
popular
in
the
West.
• Rela0onships
and
Events
–
popular
in
the
East.
• Objects
and
Events
–
seems
to
be
the
new
trend.
• The
Web
has
re-‐emphasized
the
importance
of
every
object
and
event
being
connected
to
others
-‐-‐
East
Meets
West.
Geography
of
Thought
by
Richard
NisbeB
17. • Take
place
in
the
real
world.
• Captured
using
different
sensory
mechanism.
– Each
sensor
captures
only
a
limited
aspect
of
the
event.
• Can
be
used
to
bridge
the
seman1c
gap.
18. Events:
Types
and
Granulari1es
• Conferences
– Days
• Sessions
– Talks
» Purpose
of
the
talk
• Wedding
• An
Earthquake
• The
Big
Bang
• World
Wide
Web
• Yahoo:
Winter
School
2012
• Me
– My
Birth,
– Being
here,
and
– Dying
in
100
years.
19. People
Things
Places
Time
Experiences
Events
E
by
Westerman
and
Jain
E*
by
Gupta
and
Jain
21.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
23. Connec4ng
People
to
Resources
effec4vely,
efficiently,
and
promptly
in
given
situa4ons.
24. • Minimize
hunger
in
the
world.
• Maximize
female
educa1on
in
India.
• Minimize
‘deaths’
in
the
coming
hurricane
in
Florida.
• Minimize
work-‐hours
lost
in
traffic
during
week
days
in
Bangalore.
25. • System:
–
A
set
of
diverse
parts
forming
a
whole.
– Parts
are
put
together
with
a
common
objec1ve/
purpose.
• Each
part
could
be
considered
a
system.
• Each
part
plays
a
role
towards
the
system
objec1ve.
• Designing
the
informa1on
flow
among
parts
is
essen1al
to
make
a
system
work
apprpriately.
26. • A
social
system
is
composed
of
persons
or
groups
who
share
a
common
objec1ve.
• An
individual
objec1ve
is
usually
a
part
of
the
group’s
objec1ve.
27. • Persons
• Families
• Organiza1ons
• Communi1es:
City,
State,
Country
• Socie1es
• Cultures
28. • Top
Down:
– The
social
system
determines
its
parts.
– People’s
behavior
determined
by
society.
• BoBom
Up:
– The
Society
is
the
sum
of
its
indivduals
– Individual
ac1ons
determine
the
character
of
the
society.
29. • Each
social
en1ty
is
a
holon.
• Holon:
Each
en1ty
is
simultaneously
a
part
and
a
whole.
• A
social
component
is
made
up
of
parts
and
at
the
same
1me
maybe
part
of
some
larger
whole.
• Any
system
is
by
defini1on
both
part
and
whole.
30. • The
primary
‘currency’
of
a
social
system
is
informa1on.
• System
behavior
can
be
understood
as
the
movement
of
informa1on:
– Within
a
system
– Between
the
system
and
its
environment
• Informa1on
is
used
to
sense
as
well
as
to
control
or
act.
31.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
32. • Systems
that
perceive,
reason,
learn,
and
act
intelligently.
• Adaptability
to
varying
environmental
situa1ons
is
a
key
element
of
intelligent
systems
33. • Social
systems
that
perceive,
reason,
learn,
and
act
intelligently.
• What
does
‘perceive’,
‘reason’,
‘learn’,
and
‘act’
mean
in
the
context
of
social
systems?
34.
35.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Social
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
36.
37.
38.
39.
40.
41. • Desired
state
(Goal)
• System
model
and
Control
Signal
(Ac0ons)
• Current
State
(for
Feedback)
42. bserve d
State
O
Fe edback
Observa0ons
Control
Signals
Events
Real
World
48. Connecting Information
People
Aggregation Situation Alerts
and Detection
CompositionAnd
Queries
Resources
12/5/12
48
49.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Systems
• Situa0on
Recogni0on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
50. Connec4ng
People
to
Resources
effec4vely,
efficiently,
and
promptly
in
given
situa4ons.
51. • rela1ve
posi1on
or
combina1on
of
circumstances
at
a
certain
moment.
• The
combina1on
of
circumstances
at
a
given
moment;
a
state
of
affairs.
52. • Situa1on
awareness,
or
SA,
is
the
percep1on
of
environmental
elements
within
a
volume
of
1me
and
space,
the
comprehension
of
their
meaning,
and
the
projec1on
of
their
status
in
the
near
future.
• What
is
happening
around
you
to
understand
how
informa1on,
events,
and
your
own
ac1ons
will
impact
your
goals
and
objec1ves,
both
now
and
in
the
near
future.
53. • Example
1:
– A
person
shou1ng.
– 1000
people
shou1ng.
• In
a
contained
building
• In
main
parts
of
a
city
• Example
2:
– One
person
complaining
about
flu.
– Many
people
from
different
areas
of
a
country
complaining
about
flu.
54. Facebook
and
TwiBer
(now
GOOGLE
+)
Have
been
repor0ng
events
as
micro-‐blogs
Massive
collec1on
of
events.
60. • Given
a
plethora
of
event
data.
How
can
we:
– Disambiguate
relevant
and
irrelevant
events?
– Combine
events
into
meaningful
representa1ons
?
– Allow
inference
and
cascading
effects?
– Support
different
interpreta1ons
based
on
applica1on
domain?
– Support
Control
&
decision
making?
61. 1. Inherent
support
for
event-‐based
(temporal)
reasoning
2. The
ability
of
the
controller
to
reason
based
on
symbols
(rather
than
just
signals)
3. Explicit
inclusion
of
domain
seman1cs
(to
support
mul1ple
applica1ons)
63.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Systems
• Situa1on
Recogni1on
• Concept
recogni0ons
• Personalized
Situa1ons
• EventShop
65. Situa0on
2010
Events
Data
Type
1986
Objects
1963
Speech
1962
Character
1959
1950
2000
Time
Line
66. Loca1on
Scenes
Environm Trajectories
Situa1ons
aware
ents
Single
Media
Loca1on
Visual
Real
world
Visual
Objects
Objects
Ac1vi1es
Events
unaware
Sta1c
Dynamic
SPACE
TIME
Data
=
Text
or
Images
or
Video
66
68. Heterogeneous
Media
Loca1on
Environm
Situa1ons
aware
ents
Loca1on
Real
world
Objects
Ac1vi1es
unaware
Sta1c
Dynamic
SPACE
TIME
Data
is
just
Data.
Meta-‐data
is
also
data.
Caste
system
does
not
exist
here.
Medium
and
sources
do
not
maOer.
68
69.
• Introduc1on
• Social
Systems
• Real
Time
Social
Systems
• Designing
Real
Time
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa0ons
• EventShop
70. A)
Situa0on
Modeling
B)
Situa0on
Recogni0on
C)
Visualiza0on,
Personaliza0on,
and
Alerts
i)
Visualiza1on
C1
…
⊕
v2
v3
Personal
context
ii)
Personaliza1on
Personali
v5
v6
zed
STT
Stream
situa1on
Available
resources
Emage
iii)
Alerts
Situa1on
70
73. Aggrega1on,
Opera1ons
Alert
level
=
High
Date:
3rd
Jun,
2011
STT
data
Situa1on
Detec1on
User-‐Feedback
1)
Classifica1on
Tweet:
‘Please
visit
Dr.
Cureit
at
‘Urrgh…
sinus’
2)
Control
ac1on
4th
St
immediately’
Loc:
NYC,
Date:
3rd
Jun,
2011
Theme:
Allergy
73
74.
75.
• Introduc1on
• Social
Systems
• Intelligent
Social
Systems
• Designing
Intelligent
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
76. • E-‐mage
– Visualiza1on
– Intui1ve
query
and
mental
model
– Common
spa1o
temporal
data
representa1on
– Data
analysis
using
media
processing
operators
(e.g.
segmenta1on,
background
subtrac1on,
convolu1on)
76
80. Retail
Store
Loca0ons
Net
Catchment
area
Proprietary
and
Confiden1al,
Not
For
12/5/12
80
Distribu1on
81. • Humans
as
sensors
• Space
+
Time
as
fundamental
axes
• Real
0me
situa0on
evalua0on
(E-‐mage
Streams)
(a) Pollen levels (Source: Visual) (b) Census data (Source: text file) (c) Reports on ‘Hurricanes’ (source: Twitter stream)
d) Cloud cover (Source: Satellite imagery) (e) Predicted hurricane path (source: KML) (f) Open shelters coverage(Source: KML)
81
82. • Help
domain
experts
externalize
their
internal
models
of
situa1ons
of
interest
e.g.
epidemic.
• Building
blocks:
– Operators
– Operands
• Wizard:
– A
prescrip1ve
approach
for
modeling
situa1ons
using
the
operators
and
operands
Singh,
Gao,
Jain:
Situa:on
recogni:on:
An
evolving
problem
for
heterogeneous
dynamic
big
82
mul:media
data,
ACM
Mul0media
‘12.
83. Knowledge
or
data
driven
building
blocks
Growth
rate
(Flu
reports)
Feature
TwiBer-‐Flu
Data
source
-‐Emage
(#Reports)
Representa1on
level
Thresholds
(0,
50)
Meta-‐data
83
88. Suppor1ng
Operator
Type
Data
parameter(s)
Output
1)
Data
into
right
representa1on
Transform
…
Spa1o-‐temporal
window
Filter
+
Mask
Aggregate
+
2)
Analyze
data
to
Classifica1on
derive
features
Classifica1on
method
Characteriza1on
Property
Growth
Rate
required
=
125%
PaBern
Matching
+
PaBern 72%
{Features}
3)
Use
features
to
Learn
f
Learning
f
evaluate
situa1ons
{Situa1on}
method
88
90. • Select
E-‐mages
of
US
for
theme
‘Obama’.
– ∏spa1al(region=[24,-‐125],[24,-‐65])
(TEStheme=Obama)
• Iden1fy
3
clusters
for
each
E-‐mage
above.
– γkmeans(3)
(∏spa1al(region=[24,-‐125],[24,-‐65])(TEStheme=Obama))
• Show
me
the
speed
for
each
cluster
of
‘Katrina’
e-‐
mages
(
– @speed @epicenter (γkmeans(n=3)
(∏spa1al(region=[24,-‐125],[24,-‐65])
(TEStheme=Katrina))))
• How
similar
is
paBern
above
to
‘exponen1al
increase’?
– ψexp-‐increase(@speed(@epicenter(γkmeans(n=3)
(∏spa1al(region=[24,-‐125],[24,-‐65])
(TEStheme=Katrina))))
90
91. Personalized
situa0on:
An
ac4onable
integra4on
of
a
user's
personal
context
with
surrounding
spa4otemporal
situa4on.
1)
Macro
situa0on
Macro
Personal
2)
Personalized
data-‐sources
Context
situa0on
Profile
+
Preferences
3)
Personalized
alerts
User
Available
data
resources
Resource
data
IF
person
ui
<is-‐in>
(PSj)
THEN
<connect-‐to>
rk
91
94. Billions
of
data
sources.
Selec0ng
and
combining
appropriate
sources
to
detect
situa0ons.
Interac0ons
with
different
types
of
Users
Decision
Makers
Individuals
12/5/12
94
96. Front
End
GUI
New New E-‐mage Alert
Data Query Stream Request
Source
Back
End
Controller
E-‐mage
Stream
Personalized
Registered Stream
Query
Processor
Queries Alert
Unit
E-‐mage
Stream User
Info
Registered
Data Data
Ingestor Raw
Data
Storage
Sources
API
Calls Raw
Spatial
Data
Stream
Data
Cloud
12/5/12
96
99. Macro
situa0on
Alert
Level=High
Date=12/09/10
Micro
event
Situa0onal
controller
Control
Ac0on
e.g.
“Arrgggh,
I
“Please
visit
have
a
sore
throat”
• Goal
nearest
CDC
(Loc=New
York,
• Macro
Situa1on
center
at
4th
St
Date=12/09/10)
• Rules
immediately”
Level
1
personal
threat
+
Level
3
Macro
threat
-‐>
Immediate
ac0on
12/5/12
99
100. • What
personal
informa1on
can
be
shared?
• How
should
it
be
shared
to
benefit
the
user?
• Developing
an
architecture
for
personal
informa1on
management.
110. Outline
• Introduc1on
• Social
Systems
• Real
Time
Social
Systems
• Designing
Real
Time
Systems
• Situa1on
Recogni1on
• Concept
recogni1ons
• Personalized
Situa1ons
• EventShop
• Going
Forward
111. • Social
observa1ons
are
now
possible
with
liBle
latency.
• Now
possible
to
design
social
systems
with
feedback.
• Situa1on
Recogni1on
and
Need-‐Availability
iden1fica1on
of
resources
becomes
a
major
challenge.
•
EventShop
is
a
step
in
the
direc1on
of
implemen1ng
Social
Life
Networks.
112. Useful
Links
• Demo:
– hBp://auge.ics.uci.edu/eventshop
• Data
Defini1on
Language
Schema
– hBp://auge.ics.uci.edu/eventshop/documents/
EventShop_DDL_Schema
• Query
Language
Schema
– hBp://auge.ics.uci.edu/eventshop/documents/
EventShop_QL_Schema
• Example
Query
in
JSON
– hBp://auge.ics.uci.edu/eventshop/documents/
EventShop_Example_Query
11/28/2012
112
113. Thanks
for
your
1me
and
aBen1on.
For
ques1ons:
jain@ics.uci.edu