"Computing for Human Experience: Semantics empowered Cyber-Physical, Social and Ubiquitous Computing beyond the Web" Keynote at On the Move Federated Conferences, Crete, Greece, October 18, 2011.
http://www.onthemove-conferences.org/
Details: http://wiki.knoesis.org/index.php/Computi
Computing for Human Experience [v4]: Keynote @ OnTheMove Federated Conferences
1. Computing for Human Experience
Keynote at On-the-Move Federated Conference, October 2011:
http://www.onthemove-conferences.org/
Amit Sheth
Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing
Wright State University, Dayton, OH, USA
Special thanks & contributions: Cory Henson
1
2. Our 1 world
with 7 billion people
living 1 experience at a time
2
4. Today, technology increasingly engages
individuals, society and humanity with …
1-2 billion
computers
5 billion 40+ billion
mobile phones mobile sensors
http://www.gartner.com/it/page.jsp?id=703807 4
5. With constant connectivity enabled through global networks
1-2 billion
computers
5 billion 40+ billion
mobile phones mobile sensors
5
6. So, how can we leverage this tech to improve our experiences
without losing ourselves in the process?
6
8. From machine-centric to human-centric design
Machines to accommodate our experiences,
as opposed to the other way around
Computing to liberate
8
9. To accomplish this requires a fundamental
shift in how we interact and communicate
with computational machines
We must take a more holistic view of computation,
as a shared universe, populated by people and machines
working in harmony to achieve our highest aspirations
9
10. Ubiquitous Computing – M. Weiser
We have caught glimpses of this vision …
Memex– V. Bush
10
11. The ways in which technology and humans
interact are fundamentally changing
In turn, our (human) experiences are changing
Our activities, decisions, thoughts, and feelings
are affected by the ubiquitous integration of
technology into the fabric of our lives
Let this affect be positive
11
12. Computing for Human Experience
AmitSheth
Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing
Wright State University, Dayton, OH, USA
Special thanks & contributions: Cory Henson
12
13. Alan Smith Vinh Michael HemantP
Nguyen Cooney urohit Matthan Sink
Sujan
Perera
Wenbo
Wang
Pramod Koneru
Cory Henson
Amit Sheth
Maryam Panahiazar
Kalpa
Gunaratna
AshutoshJadhav
Sanjaya
Wijeratne
Pramod Prateek PavanKapanip Delroy
Sarasi Lalithsena Ajith
Anantharam Jain athi Lu Chen Cameron
Ranabahu
14. CHE is an approach to improving the human
condition through computational means, and
with minimal burden
This may be achieved through the contextual
assistance, augmentation, and absolution of human
capabilities
Human capabilities such as
sensing, perception, attention, memory,
decision making, control, etc.
14
16. A cross-country flight from New York to Los Angeles on a
Boeing 737 plane generates a massive 240 terabytes of data
- GigaOmni Media
16
17. But, how much data is generated regarding the
health and well-being of the pilot or passengers?
zero, none, zilch!
17
18. Image the ability to monitor and control our health
with the same care and precisionthat goes into the 737.
And not just providing
doctors with such
control, but you and me.
18
19. Health information is now available from multiple sources
• medical records
• background knowledge
• social networks
• personal observations
• sensors
• etc.
19
20. Sensors, actuators, and mobile computing are playing an
increasingly important role in providing data for early phases of
the health-care life-cycle
This represents a fundamental shift:
• people are now empowered to monitor and manage their own health;
• and doctors are given access to more data about their patients
20
22. What is needed is a more intuitive and
intelligent representation of our health.
Personal Health Dashboard
Health Metrics with Meaning
Image: http://bit.ly/lV2V73 22
24. Key Enablers of CHE
• Integration of heterogeneous, multimodal data
• Bridging the physical-cyber-social divide
• Elevating abstractions that machines and people understand
• Semantics at an extraordinary Scale
These enablers are brought together through
Semantic Web technologies
24
28. Integration of heterogeneous, multimodal data
Background Knowledge: ontologies, knowledge bases, LOD,
databases, etc.
Social/Community Data:social network data, wisdom of the
crowds, etc.
Sensor Data: observations from machine sensors, citizen sensors (i.e.,
patients, doctors), laboratory experiments, etc.
Personal Context: location, schedule, items (e.g., accessible sensors),
etc.
Personal Medical History: Electronic Medical Records, Personal
Health Records, Patient Visit Records, etc.
28
29. What is Social Media?
Communications using online technologies
to share opinions, insights, experiences and
perspectives with each other.
29
30. Popular types of Healthcare Social Media
Blogs – DiabetesMine, HealthMatters, WebMD, NYT HealthBlog, etc.
Microblogs –Livestrong, Stupid Cancer, etc.
Social Networks –
OrganizedWisdom, PatientsLikeMe, DailyStrength, NursesRecommendDoctors, CureTog
ether, etc.
Podcasts – John Hopkins Medical Podcasts, Mayo Clinic, etc.
Forums – Revolution Health Groups, Google Health Groups, etc.
30
31. HCPs aren’t waiting to be detailed, they’re turning to the
social web to educate themselves
60% ofof physicians either or are interested in using social networksnetworks
60% physicians either use use or are interested in using social
112,000 docs
talk to each
other on Sermo. This doc-to-doc
blogger has
53,000 readers
this month +
20,000 Twitter
followers
65% of docs plan to use
social media for
professional development
Manhattan Research 2009, 2010
Sermo,com
Compete.com
http://www.slideshare.net/IQLab/social-media-101-for-pharma-3494462 31
32. People are turning to each other online to understand their health
http://www.slideshare.net/IQLab/social-media-101-for-pharma-3494462 32
33. 83% of online adults search for health information
http://www.slideshare.net/IQLab/social-media-101-for-pharma-3494462 33
34. 83% of online adults search for health information
60% of them look for the experience of “someone like me”
http://www.slideshare.net/IQLab/social-media-101-for-pharma-3494462 34
35. "I don't know, but I can try to find out" is the
default setting for people with health questions.
"I know, and I want to share my knowledge"
is the leading edge of health care.
Savannah Fox, The Social Life of Health Information, Pew Internet Report, May 12, 2011.
Available at http://www.pewinternet.org/Reports/2011/Social-Life-of-Health-Info.aspx 35
36. People-Content-Network Analysis
Intra Community Activity and connectivity
– How well connected are individual nodes (People)
– What keeps them strongly connected over time
(Relationship types - Knowledge of Content)
Will the two communities coordinate well
during an event- crisis or disaster?
• Interplay between all three dimensions –
P, C, N
Inter-Community Connectivity
• Any bridges to connect to the other community?
(People)
• Any Similarity in actions with the other community
Image: http://themelis-cuiper.com (Can Content help?)
For more info: http://www.slideshare.net/knoesis/understanding-usercommunity-engagement-
by-multifaceted-features-a-case-study-on-twitter 36
37. People-Content-Network Analysis
Event oriented
Community
External
Knowledge bases
Dynamic
Domain
Model SEMANTIC
for the ASSOCIATION TO Social Network
event UNDERSTAND
ENHANCED
ENGAGEMENT LEVEL
User
Mined User Interests Profiles
and User Types
37
38. Bridging the physical-cyber-social divide
Computation is no longer confined to pure symbol manipulation. The previously strict
relations between the digital and physical world are blurring.
38
39. Bridging the physical-cyber divide
Psyleron’s Mind-Lamp (Princeton U),
connections between the mind and the
physical world.
MIT’s Fluid Interface Group: wearable
device with a projector for deep
interactions with the environment
Neuro Sky's mind-controlled headset to
play a video game.
39
40. Bridging the physical-cyber-social divide
FitBit Community allows the
automated collection and
sharing of health-related data,
goals, and achievements
Foursquare is an online application which
integrates a persons physical location and Community of enthusiasts that share experiences of
social network. self-tracking and measurement.
40
45. The design and building of physical-cyber-social systems requires effective
conceptualization and communication between people and machines.
To reach this vision requires advancement in the area of machine perception,
enabling machines the ability to abstract over low-level observations.
45
46. Abstraction provides the ability to interpret and synthesize information in a way that
affords effective understanding and communication of ideas, feelings, perceptions, etc.
between machines and people.
Abstraction
46
47. People are excellent at abstraction; of
sensing and interpreting stimuli to
understand and interact with the world.
The process of interpreting stimuli is called perception;
and studying this extraordinary human capability can lead
to insights for developing effective machine perception.
47
48. Sensor
“real-world”
observation
Physical
Social
Cyber
perception
conceptualization of
“real-world” Sensor Data /
Social Data
48
49. Both people and machines are capable of observing qualities, such as redness.
observes
Observer Quality
* Formally described in a sensor/observation ontology
49
50. Sensor and Sensor Network (SSN) Ontology
http://www.w3.org/2005/Incubator/ssn/XGR-ssn/
50
51. The ability to perceive is afforded through the use of background
knowledge, relating observable qualities to entities in the world.
Quality
* Formally described in
inheres in
domain ontologies
(and knowledge bases)
Entity
51
53. With the help of sophisticated inference, both people and machines are
also capable of perceiving entities, such as apples.
perceives
Perceiver Entity
• the ability to degrade gracefully with incomplete information
• the ability to minimize explanations based on new information
• the ability to reason over data on the Web
•fast (tractable)
53
54. Abductive Logic Deductive Logic (e.g., OWL)
high complexity (relatively) low complexity
minimize
explanations tractable
degrade gracefully Web reasoning
Perceptual Inference
(i.e., abstraction)
Cory Henson, KrishnaprasadThirunarayan, AmitSheth, Pascal Hitzler. Representation of Parsimonious Covering
Theory in OWL-DL. Proceedings of the 8th International Workshop on OWL: Experiences and Directions (OWLED
2011), San Francisco, CA, USA, June 5-6, 2011.
54
55. Parsimonious Covering Theory
disorder causes manifestation
m1
• Goal is to account for observed symptoms d1
with plausible explanatory hypotheses m2
(abductive logic) d2
m3
d3
m4
• Driven by background knowledge modeled explanation
observations
as a bipartite graph causally linking
disorders to manifestations
YunPeng, James A. Reggia, "Abductive Inference Models for Diagnostic Problem-Solving" 55
56. PCT Parsimonious Cover
• coverage: an explanation is a cover if, for each observation, there is
a causal relation from a disorder contained in the explanation to the
observation
• parsimony: an explanation is parsimonious, or best, if it matches
some criteria of suitability (i.e., single disorder assumption)
56
57. Convert PCT to OWL
Given
PCT problem P is a 4-tuple ⟨D, M, C, Γ⟩
• D is a finite set of disorders
• M is a finite set of manifestations
• C is the causation function [C : D ⟶ Powerset(M)]
• Γ is the set of observations [Γ ⊆ M ]
Δ is a valid explanation (i.e., is a parsimonious cover)
Goal
Translate P into OWL, o(P), such that o(P) ⊧ Δ
Cory Henson, KrishnaprasadThirunarayan, AmitSheth, Pascal Hitzler. Representation of Parsimonious Covering
Theory in OWL-DL. Proceedings of the 8th International Workshop on OWL: Experiences and Directions (OWLED
2011), San Francisco, CA, USA, June 5-6, 2011.
57
58. disorders (D)
PCT Background for all d ∈ D, write d rdf:type Disorder
Knowledge in OWL ex: flu rdf:type Disorder
cold rdf:type Disorder
disorder causes manifestation
fever
manifestations (M)
headache
extreme exhaustion for all m ∈ M, write m rdf:type Manifestation
flu
severe ache and pain ex: fever rdf:type Manifestation
mild ache and pain
stuffy nose
headache rdf:type Manifestation …
sneezing
cold
sore throat
causes relations (C)
severe cough
mild cough for all (d, m) ∈ C, write d causes m
ex: flu causes fever
flu causes headache …
Cory Henson, KrishnaprasadThirunarayan, AmitSheth, Pascal Hitzler. Representation of Parsimonious Covering
Theory in OWL-DL. Proceedings of the 8th International Workshop on OWL: Experiences and Directions (OWLED
2011), San Francisco, CA, USA, June 5-6, 2011.
58
59. observations (Γ)
for mi∈ Γ, i =1 … n, write
PCT Observations and
Explanation owl:equivalentClass
Explanations in OWL causes value m1 and … causes value mn
ex: Explanation owl:equivalentClass
causes value sneezing and
causes value sore-throat
and
causes value mild-cough
explanation (Δ)
Δrdf:type Explanation, is deduced
ex: cold rdf:type Explanation
flu rdf:type Explanation
Cory Henson, KrishnaprasadThirunarayan, AmitSheth, Pascal Hitzler. Representation of Parsimonious Covering
Theory in OWL-DL. Proceedings of the 8th International Workshop on OWL: Experiences and Directions (OWLED
2011), San Francisco, CA, USA, June 5-6, 2011.
59
60. The ability to perceive efficiently is afforded through the cyclical
exchange of information between observers and perceivers.
Observer
sends sends
observation focus Traditionally called the
Perception Cycle
(or Active Perception)
Perceiver
60
62. Cognitive Theories of Perception (timeline)
1970’s – Perception is an active, cyclical process of exploration and
interpretation. - Nessier’s Perception Cycle
1980’s – The perception cycle is driven by background knowledge in
order to generate and test hypotheses. - Richard Gregory (optical illusions)
1990’s – In order to effectively test hypotheses, some observations are
more informative than others. - Norwich’s Entropy Theory of Perception
62
63. Key Insights
• Background knowledge plays a crucial role in perception; what we know (or
think we know/believe) influences our perception of the world.
• Semantics will allow us to realize computational models of perception based
on background knowledge.
Contemporary Issues
• Internet/Web expands our background knowledge to a global scope; thus
our perception is global in scope
• Social networks influence our knowledge and beliefs, thus influencing our
perception
63
64. Integrated together, we have an general model – capable of abstraction –
relating observers, perceivers, and background knowledge.
observes
Observer Quality
sends sends
observation inheres in
focus
perceives
Perceiver Entity
64
65. Modeled in set-theoretic notation with components
mapped to Parsimonious Covering Theory and OWL
Cory Henson, KrishnaprasadThirunarayan, AmitSheth. An Ontological Approach to Focusing Attention and Enhancing
Machine Perception on the Web. Applied Ontology, 2011 (accepted).
65
67. Weather Application
Weather Application
Detection of events, such as blizzards, from weather
station observations on LinkedSensorData
Demos: Real-Time Feature Streams 67
68. SECURE: Semantics Empowered Rescue Application
Weather Environment
Rescue robots detect different types of fires, which may require different
methods/tools to extinguish, and relays this knowledge to first responders.
Demo:SECURE: Semantics Empowered Rescue Environment 68
69. Healthcare Application
Weather Application
EMR
Detection of errors in Electronic Medical Records and
missing knowledge in a cardiology domain model
69
70. Healthcare Application
Weather Application
EMR: "Her prognosis is poor both short term and long term, however, we
will do everything possible to keep her alive and battle this infection."
without IntellegO
with IntellegO
Problem Problem
SNM:40733004_infection SNM:68566005_infection_urinary_tract
A syntax based NLP extractor By utilizing IntellegO and cardiology
(such as Medlee) can extract background knowledge, we can more
this term and annotate accurately annotate the term as
asSNM:40733004_infection SNM:68566005_infection_urinary_tract
70
71. Healthcare Application
Weather Application
EMR: ”The patient is to receive 2 fluid buloses."
without IntellegO with IntellegO
Problem Treatment
SNM:32457005_body_fluid Fluid is part of buloses treatment, not a problem
A syntax based NLP extractor By utilizing IntellegO and cardiology
(such as Medlee) can extract background knowledge, we can determine
this term and annotate that this is an incorrect annotation.
asSNM:32457005_body_fluid
71
72. Semantic Scalability
In 2008, the rate of data generation surpassed storage capacity. With 7 billion people, and a
growing number of sensors, how can such a such a system scale? By shining a light on
relevant human experience, supported by knowledge, while dimming the minutia of data.
http://gigaom.com/cloud/sensor-networks-top-social-networks-for-big-data-2 72
73. Semantic Scalability
It is clear that purely syntax-based solutions will not scale
ex. – keyword based search/index, non-textual data, multimodal data for an event
73
74. Semantic Scalability
Path to Web scale semantics
1. Focusing attention on important information and ignoring irrelevant data
2. Converting low-level data (observations) to high-level knowledge (abstractions)
3. Utilizing CHE technology to more evenly distribute responsibility and activities
among people and machines
74
75. 1. Focusing attention on important
information and ignoring irrelevant data
We were able to demonstrate 50% savings in sensing
resource requirements during the detection of a blizzard.
Cory Henson, KrishnaprasadThirunarayan, AmitSheth. An Ontological Approach to Focusing Attention and Enhancing
Machine Perception on the Web. Applied Ontology, 2012. (accepted)
75
76. 2. Converting low-level data to
high-level knowledge
(observations to abstractions)
Experiment – during a blizzard, we utilized Intelleg0 to collect
and analyze over 110,000 sensor observations, from:
• 800 weather stations (~5 sensors per station)
• across 5 states (Utah, Nevada, Colorado, Wyoming, and Idaho)
• for 6 days (April 1 – 6, 2003)
Cory Henson, KrishnaprasadThirunarayan, AmitSheth. An Ontological Approach to Focusing Attention and Enhancing
Machine Perception on the Web. Applied Ontology, 2012. (accepted)
76
77. 2. Converting low-level data to
high-level knowledge
(observations to abstractions)
We were able to demonstrate an order of magnitude resource
savings between storing observations vs. relevant abstractions
Cory Henson, KrishnaprasadThirunarayan, AmitSheth. An Ontological Approach to Focusing Attention and Enhancing
Machine Perception on the Web. Applied Ontology, 2012. (accepted)
77
78. 2. Converting low-level data to
high-level knowledge
(observations to abstractions)
While this is a good result, the benefit provided for a single
person – a single experience – is far more dramatic.
Cory Henson, KrishnaprasadThirunarayan, AmitSheth. An Ontological Approach to Focusing Attention and Enhancing
Machine Perception on the Web. Applied Ontology, 2012. (accepted)
78
79. 3. Utilizing CHE technology to more evenly distribute
responsibility and activities among people and machines
There are almost 7 billion people on earth, and only ~10-15 million doctors (~700:1 - 467:1)
79
80. These doctors are severely overburdened,
answering less than 60% of questions posed by
patients regarding their health and well being
Providing people with the tools to monitor and manage their own health will
dramatically reduce the burden on doctors, and improve the health of the people
80
81. The health-care ecosystem of the future includes
machines, people, and social networks
continuously, ubiquitously, and unobtrusively
monitoring and managing our health
81
82. CHE approach to Health Care
Continuous Monitoring Personal Assessment Medical Service
1 2 3
Auxiliary Information – background knowledge, social/community support,
personal context, personal medical history
82
83. Continuous Monitoring Phase
Monitoring health metrics and vital signs utilizing unobtrusive body sensors
Continuously collecting information, watching for worrisome symptoms
83
84. Personal Assessment Phase
Assessment of symptoms from personal observation and/or health sensors
available at home.
Utilizing background knowledge, personal medical history, and current
sensor data to formulate and ask specific questions of patient that will aid in
explaining symptoms.
84
85. Medical Services Phase
Assessment of symptoms gathered from continuous and personal
phases, with additional sophisticated equipment, advanced treatment, and
specialized medical knowledge not previously available.
Utilizing background knowledge, personal medical history, and current
sensor data to formulate a diagnosis.
85
86. CHE approach to Health Care
Continuous Personal Medical
Symptoms/ Symptoms/
Explanations Explanations
access & access & access &
update (PHR) update (PHR) update (EMR)
Auxiliary Information
Background Knowledge Social/Community Support Personal Context Personal Medical History
(e.g., available sensors, (e.g., Electronic Medical Record,
(e.g., Ontologies, (e.g., Patient Network,
location, schedule) genomic sequence)
Knowledgebases) crowd sourcing)
86
87. Continuous Monitoring Phase: Example
Observed Symptoms Possible Explanations
• Abnormal heart rate • Panic Disorder
• Clammy skin • Hypoglycemia
• Hyperthyroidism
• Heart Attack
• Septic Shock
Electronic Medical Record Health Alert
• Patient has history of Heart Disease • Check phone for instructions
87
88. Continuous Phase Technology
Basis is a wrist-watch that also monitors pulse rate, movement, temperature, and
galvanic skin response.
88
89. Continuous Phase Technology
Fitbit Tracker uses a MEMS 3-axis accelerometer that measures your motion
patterns to tell you your calories burned, steps taken, distance traveled, and sleep
quality.
89
90. Personal Assessment Phase: Example
Are you feeling lightheaded?
yes
Are you have trouble taking deep breaths? Observed Symptoms Possible Explanations
• Abnormal heart rate • Panic Disorder
yes •
• Clammy skin Hypoglycemia
• Lightheaded • Hyperthyroidism
Do you have low blood pressure when • Trouble breathing • Heart Attack
standing? •
• Low blood pressure Septic Shock
yes
Have you taken your Methimazole
medication?
no
Electronic Medical Record Health Alert
• Patient has history of Hyperthyroidism 1. Take medication: Methimazole
• Patient has prescription for Methimazole 2. See doctor: how about Tues. @ 11am?
90
91. Personal Phase Technology
Lark is a sleep sensor that monitors circadian rhythms and functions as an "un-
alarm," vibrating to wake you at a point in your sleep cycle when you feel alert, not
groggy.
91
92. Personal Phase Technology
Instant Heart Rate takes your pulse when you place your finger over your phone’s
camera lens. The app uses light from the camera flash to detect color changes caused
by blood moving through your finger.
92
93. Personal Phase Technology
Telcare makes a blood glucose meter (right) for diabetics that broadcasts readings to
a mobile-phone app (center) where patients can see results and set goals.
93
94. Personal Phase Technology
iBGStar is a plug-in glucose meter for the iPhone, developed by Sanofi-
Aventis, providing the ability for patients to monitor and manage Diabetes.
94
95. Personal Phase Technology
Withings Blood Pressure Monitor provides easy and convenient blood pressure
readings in the convenience of home.
95
96. Personal Phase Technology
WebMD provides a wealth of health information and an application to diagnose
symptoms.
96
97. Medical Services Phase: Example
Are your blood sugar levels low?
Observed Symptoms Possible Explanations
• Abnormal heart rate • Hypoglycemia
• Clammy skin • Hyperthyroidism
• Lightheaded
• Trouble breathing
• Low blood pressure
Electronic Medical Record
• Patient has history of Hyperthyroidism
• Patient has prescription for Methimazole
97
99. Medical Phase Technology
Health Guard provides a secure way to store and analyze health records for casual
browsing or emergency use (i.e., MS Health Vault records).
99
100. Medical Phase Technology
Mobile MIM gives physicians a sophisticated, hands-on mobile system for viewing
and annotating radiology images, such as CT scans.
100
101. Medical Phase Technology
Dr. Watson is a health and medical question and answering system developed by
IBM, utilizing supercomputer intelligence for medical diagnostics.
101
102. CHE approach to Health Care
Continuous Personal Medical
Symptoms/ Symptoms/
Explanations Explanations
access & access & access &
update (PHR) update (PHR) update (EMR)
Auxiliary Information
Background Knowledge Social/Community Support Personal Context Personal Medical History
(e.g., available sensors, (e.g., Electronic Medical Record,
(e.g., Ontologies, (e.g., Patient
location, schedule) genomic sequence)
Knowledgebases) Network, crowd sourcing)
102
103. CHE holds the potential to revolutionize the practice
of health-care by embracing the relationship between
ourselves, our machines, and our health
Improving the experience of health-care
improves all other experiences
103
104. Physical-Cyber-Social
Abstraction
“The most profound technologies are those that disappear.
They weave themselves into the fabric of everyday life until
they are indistinguishable from it.” – M. Weiser
Integration Scalability
104
105. Computing for Human Experience
thank you, and please visit us at
http://knoesis.org
Kno.e.sis – Ohio Center of Excellence in Knowledge-enabled Computing
Wright State University, Dayton, Ohio, USA
More: Vision Paper: Computing for Human
Experience:http://wiki.knoesis.org/index.php/Computing_For_Human_Experience
105
Imageshttp://www.cbi.umn.edu/about/babbage.htmlhttp://news.cnet.com/8301-13579_3-10141672-37.htmlhttps://fpm-www3.fpm.wisc.edu/safety/occupationalHealth/Ergonomics/LaptopErgonomics/LaptopErgonomicsQuickReference/tabid/106/Default.aspxHow many personal computers (PC) in the world -- http://www.gartner.com/it/page.jsp?id=703807
Use-Case for People-Content-Network analysis - Not limited to user-community engagement, but it can answer many questions, exploiting potential of citizen-sensingPresentation- http://www.slideshare.net/knoesis/understanding-usercommunity-engagement-by-multifaceted-features-a-case-study-on-twitterReferencesH. Purohit, Y. Ruan, A. Joshi, S. Parthasarathy, A. Sheth.Understanding User-Community Engagement by Multi-faceted Features: A Case Study on Twitter, SoME 2011, Workshop on Social Media Engagement, in conjunction with WWW 2011.M. Nagarajan, H. Purohit, A. Sheth. A Qualitative Examination of Topical Tweet and Retweet Practices , 4th Int'l AAAI Conference on Weblogs and Social Media, ICWSM 2010A. Sheth, H. Purohit, A. Jadhav, P. Kapanipathi, L. Chen. Understanding Events Through Analysis Of Social Media , Kno.e.sis Technical Report, 2010
There is an event oriented community being formed during an eventDynamic domain model would give enhanced understanding of what is going on in the event, and hence, the communityCommunity has people, sub-community of users in the community will form an implicit network for ‘resource sharing’ and ‘resource provider’ How do we get to know such people? Semantic analysis of user profile description, with help of external knowledges bases to tell us, ‘what is a user interest I’ will tell us, ‘who is what’ – Blogger, news journalist, trustee, Red Cross Member etc. and also, ‘who’ is interested in ‘what’This information of mined user interests and types, from users profiles, can be leveraged to perform semantic association with dynamic domain model of the event
Single disorder assumption says that an explanation is parsimonious if it covers, or explains, all observations.Abductive reasoning is generative– which has performance hit.Single entity/abstraction can account for all properties, so we can have deductive approach for finding explanation adequate.
Encoding background knowledge in OWL (bipartite model is converted into RDF)
Can work with incomplete informationAs new observation comes, explanation is updated
20,000 weather stations (with ~5 sensors per station)Real-Time Feature Streams - live demo: http://knoesis1.wright.edu/EventStreams/ - video demo: https://skydrive.live.com/?cid=77950e284187e848&sc=photos&id=77950E284187E848%21276
Automated detection of different types of fires, which each require different extinguishing methodsYouTubeSECURE Demo: http://www.youtube.com/user/knoesisCenter?blend=1&ob=5
- Such machines (and average/everyday people) could take some responsibilities, thus freeing up a doctors time to be more productive
Body sensors: thermometer, optical blood-flow sensor, galvanic skin response monitor, accelerometerObservations: e.g., pulse-rate is high, activity is lowHistory: previous heart attack, genetic predisposition to heart diseaseAbstraction: current condition is worrisomePotential explanation for observations = …Example current technologies: quantified-self, my life bits, Basis, fitbit, etc.
Personal sensors: self, blood-pressure monitorFocus: using knowledge of domain, patient history, and current observations, ask patient pointed questions:about symptoms only they can observe, e.g., chest painabout symptoms that can be observed by sensors the patient has current access to (i.e., blood-pressure monitor at home)about taking medications that have been prescribed for known disorders (history) that are also in current explanation (diagnosis)Observations: self observations, observations from at-home sensors (e.g., blood-pressure monitor)Abstraction: continuously update explanations based on new observations (guided by asking questions through focus)Treatment: based on current explanation, prescribe treatment:tell patient to see doctor (perhaps syncing patient and doctor calendars to recommend appointment), and send information to doctortell patient to take medication, either over-the-counter or medication that has previously been prescribed to patientCurrent example technologies: WebMD, Medlineplus, etc.
Background Knowledge (i.e., Ontologies, knowledge-bases)Domain ontology (e.g., cardiology) describing medical and health knowledgeSensor ontology describing characteristics of medical equipmentPerception ontology describing process of translating observed symptoms to actionable knowledge (i.e., diagnoses, treatments)Social/Community SupportPersonal ContextPersonal Medical HistoryIndividual medical historyRecord types: EHR, PHR, PVR, etc.Systems: MS Health VaultFamily Genetic HistoryE.g., genetic predisposition to heart disease
Going through perception cycle (regular font observation: sensors can detect; bold observation: sensors cannot detect)
Background Knowledge (i.e., Ontologies, knowledge-bases)Domain ontology (e.g., cardiology) describing medical and health knowledgeSensor ontology describing characteristics of medical equipmentPerception ontology describing process of translating observed symptoms to actionable knowledge (i.e., diagnoses, treatments)Social/Community SupportPersonal ContextPersonal Medical HistoryIndividual medical historyRecord types: EHR, PHR, PVR, etc.Systems: MS Health VaultFamily Genetic HistoryE.g., genetic predisposition to heart disease