Transcript
Our Group’s Research Themes (2014)
Vision: Computing for Human Experience.
Strategy: Innovation with real-world impact.
Themes:
• Big Data/Smart Data
• Physical-Cyber Social Computing
• Semantic Web
• Semantic Social Web and Computational Social Science Semantic
Sensor Web (IoT/Web of Things)
• Personalized Digital Health
• Health Informatics, Crisis Informatics
See: http://knoesis.org/vision
A Unique approach
Aspects Shared by Our Projects: • Real-world impact through partnership/collaboration with domain
scientists/end-users.
• Often addresses important human/social/economic development
challenges.
• Real-world scale: at minimum, developing robust prototypes (open
source tools/platforms) using real-world data and knowledge; often
pursues multidisciplinary research.
• Tendency to lead to operational applications/tools/systems,
technology transfer, products, or services.
• High-risk/high-impact; avoids incremental work.
Sample Questions Driving Our Research
• Can we predict health changes/deterioration (e.g., asthma attacks) for intervention
before an episode occurs in order to prevent them altogether?
• Can we make computers recognize implicit information present in medical records in
order to provide a comprehensive data set to the models that attempt to understand
patient health status?
• Can we automatically extract highly sparse actionable intent of emerging resource
needs and availability, as well as match them during dynamic disaster response,
which aids situational awareness & coordination?
• Can we mine societal beliefs, lack of services, and laws from social media
conversations on gender-based violence to assist regional policy makers?
• Can we empower city authorities to make policy decisions by predicting impending
city issues such as traffic, crime, and pollution?
Student Success: Our most important performance measure
Exceptional publication impact: 11 out of Sheth’s 20 PhDs have 1,000+
citations each (3 over 5,000);
Average citation for Sheth’s first 18 PhDs: 1,400+
Publications in top venues and conferences:
• ACM TIST, VLDB Journal, Comp in Human Behavior, AAAI,
CSCW Journal, Web Semantics JMIR, JBI, First Monday
• ACL, ICWSM, CSCW, ISWC, SocInfo, ESWC, VLDB, WWW, AMIA
Serving on top conference PCs: WWW, ISWC, ESWC, CIKM, ICWSM,…
Major Awards: 2015-2016 George Thomas Post-Graduate Fellowship,
2015 Eric & Wendy Schmidt Data Science for Social Good Fellowship,
2015 USAID and ICT4Peace Fellowship,
2014 ITU Telecom World Young Innovators, etc.
Outstanding or Best CS Department student award: all years since 2007
Student Success: continued
Exceptional first jobs Tenure track/tenured at higher ranked departments, e.g., George Mason U., Case
Western Reserve University, North Carolina State University
Top research labs/high-tech companies: IBM Almaden Research, IBM TJWatson
Research, Amazon, CISCO
Successful entrepreneurs
Exceptional starting salaries: six figures for M.S.
Exceptional internships: IBM Research Almaden/Watson, HP Research, Samsung
Research, Bosch R&D, Mayo, NLM, PNNL, etc.
Exceptional professional services: Invited talks; PC membership of 10-20
conferences, usually involving top conferences in their fields
50% of the Sheth’s PhDs have filed at least 2 patents; 1 former PhD has filed
30+ patents.
We are well funded:
Projects with Sheth as a PI > 7.6 million
Total Kno.e.sis active funds > 8.3 million*
* Only funds where a Kno.e.sis faculty is PI or joint PI are counted (funds
with faculty as co-PI or investigator are not counted)
CityPulse: Smart Data for Smart Cities
Physical-Cyber-Social computing for SmartCity:
Semantic integration of and reasoning over IoT/sensor and social big
data to power advanced smart city applications.
Sponsor: EU City Pulse Consortium
CityPulse: Synergistic Social/Sensor/IoT
Streaming Data Analysis
City Traffic Event Extraction from Twitter Stream
Active (green) and Scheduled (yellow)
Events in San Francisco Bay Area
Heatmap of events in Bay Area
What are the events that influence traffic in a city?
What is the impact of these events on traffic delays?
Twitris
Most comprehensive, commercial grade real-time semantic
analysis of social
Media combined with domain-specific and domain
independent background knowledge and open data.
Provides campaigns that anyone can run for deep insights
and actionable information from big social and Web data.
Components used in major crisis such as JK Floods 2014,
Chennai Floods 2015, etc. [see media coverage]
Funded by: NSF SOCS, AFRL, NSF I-CORPS, NSF PF:AIR-TT
SOCS: Social Media Based Coordination
During Disasters
During the Jammu-Kashmir flood, digital volunteers used Twitris technology and
SOCS research to identify rescue requests, to redirect the Indian Army’s rescue
team, and to identify key influencers to increase public outreach.
SOCS team provided a massive (32,000+ messages) Twitter stream and its
filtering analysis via a web interface for a regional emergency preparedness
functional exercise involving public information officers (PIOs) of hospitals, fire
departments, and the local Red Cross, as well as city and county government
officials.
It is a service to mine request-offer intentions for help, and it has been open
sourced and integrated with Ushahidi’s CrisisNET - the firehose of crisis data.
http://knoesis.org/amit/media
Sponsor: NSF
SOCS: Social Media Based Coordination
During DisastersKey technical advancements:
• Technology to identify and match intent from user-generated content (UGC)
using knowledge of psycholinguistic-patterns in the classification.
• Target-specific sentiment analysis to improve semantic understanding of UGC
by leveraging domain-specific vocabularies.
• Defining a novel method of understanding group dynamics via divergence in
group discussions over time, which merges content analysis with traditional
approach of networks, employing socio-psychological theories.
• Efficient community discovery using content and networks (OSU team).
Project: http://knoesis.org/projects/socs
Real-time matching of supply
with needs during a disaster
First Monday paper
PREDOSE: PREscription Drug abuse Online
Surveillance and Epidemiology
A social media analytics semantic web platform developed to
detect emerging patterns and trends in prescription drug abuse
through automatic information extraction from unstructured text.
Funded the same week the White House announced its
initiative to curb prescription drug abuse.
Project: http://wiki.knoesis.org/index.php/PREDOSE
Partner: Center for Interventions, Treatment and Addictions Research (CITAR)
PREDOSE: Purpose
To determine user knowledge,
attitude, and behavior related to
the non-medical use of
pharmaceutical opioids (namel
buprenorphine) as discussed on
Web forums.
To determine spatio-temporal
thematic patterns and trends in
pharmaceutical opioid abuse as
discussed on Web forums.
PREDOSE: Key Technical Issues
Use of structured background knowledge to enhance:
• Entity identification and disambiguation
• Relationship extraction
• Triple Extraction
• Sentiment Analysis
• Template Patterns
• Trend Analysis
PREDOSE: Key Outcomes
• Loperamide discovery; drug users abuse imodium for self
medication from withdrawal by megadosing. Clinical side effect
(PMVT). Reported finding led to a warning to users, issued on one
of the selected sites for the study.
• Knowledge-aware search; a hybrid approach to information retrieval
using a context free grammar (CFG) to interpret data from four
diverse categories (ontology, lexicon, lexico-ontology, rules)
kHealth
With the increasing number of affordable sensors and computing
power, it is now possible to address a large number of healthcare
challenges that can have profound implications using dHealth,
mHealth, and AI technologies.
Our specific interest is in the early indicators of health change.
kHealth
Key Research & Technology:
• Knowledge Representation to describe a domain (logic
and probabilistic).
• Semantic Perception to derive abstractions from raw
data.
• Efficient execution on resource constrained devices.
kHealth
Current kHealth applications:
• Reducing re-hospitalization of ADFH patients
• Reducing asthma incidences in children*
• Improving patient care and caregiver support for
patients with dementia*
• Reducing rehospitalization of GI patients
* On-going testing and evaluation with patients under clinical care under approved IRB
Video: http://youtu.be/mATRAQ90wio
eDrug Trends
Ohio Center of Excellence in Knowledge-Enabled Computing
Principal Investigators: Prof. Amit P. Sheth, Prof. Raminta Daniulaityte
Co-Investigators: Robert Carlson, Krishnaprasad Thirunarayan, Ramzi Nahhas,
Silvia Martins (Columbia), Edward W. Boyer (U. Mass.)
PhD Students: Farahnaz Golroo, Sanjaya Wijeratne, Lu Chen, Adarsh Alex
MS Student: Adarsh Alex
Postdoctoral Researcher: Francois Lamy
Software Engineer: Gary Smith
NIH Award#: 5 R01 DA039454-02
Trending: Social media analysis to monitor cannabis and synthetic
cannabinoid use
Timeline: 15 Sep. 2014 - 14 Sep. 2018
Award Amount: $1,689,019 + $162,505
eDrugTrends
Key questions addressed in eDrugTrends:
• How to identify and compare trends in knowledge, attitudes, and
behaviors related to cannabis and synthetic cannabinoid use
across U.S. regions with different cannabis legalization policies
using Twitter and Web forum data.
• How to identify key influencers (opinion leaders) in cannabis and
synthetic cannabinoid-related discussions on social media.
Partner: Center for Interventions, Treatment and
Addictions Research (CITAR)
Context-Aware Harassment
Detection on Social MediaPrincipal Investigators: Prof. Amit P. Sheth
Co-Investigators: Valerie Shalin, Krishnaprasad Thirunarayan
Other Faculty: Debra Steele-Johnson, Dr. Jack L. Dustin
PhD Students: Lu Chen, Wenbo Wang, Monireh Ebrahimi, Kathleen Renee Wylds
MS Students: Pranav Karan, Rajeshwari Kandakatla
Collaboration with Beavercreek High School
Ohio Center of Excellence in Knowledge-Enabled Computing
NSF Award#: CNS 1513721
TWC SBE: Medium: Context-Aware Harassment Detection on Social Media
Timeline: 01 Sep. 2015 - 31 Aug. 2018
Award Amount: $925,104 + $16,000 (REU)
Social and Physical Sensing Enabled Decision Support for
Disaster Management and Response
Principal Investigators: Prof. Amit P. Sheth, Prof. Srinivasan Parthasarathy (OSU)
Co-Principal Investigators: Densheng Liu (OSU), Ethan Kubatko (OSU), Valerie Shalin,
Krishnaprasad Thirunarayan
PhD Students: Sarasi Lalithsena, Pavan Kapanipathi, Hussein Olimat
MS Student: Siva Kumar
Postdoctoral Researcher: Tanvi Banerjee
Ohio Center of Excellence in Knowledge-Enabled Computing
NSF Award#: EAR 1520870
Hazards SEES: Social and Physical Sensing Enabled Decision Support for
Disaster Management and Response
Timeline: 01 Jul. 2015 - 31 Jul. 2019
Award Amount: $1,975,000 (WSU: $787,500)
Modeling Social Behavior for
Healthcare Utilization in DepressionPrincipal Investigators: Prof. Amit P. Sheth, Prof. Jyotishman Pathak (Cornell)
Co-Investigators: Krishnaprasad Thirunarayan, Tanvi Banerjee, William V. Bobo (Mayo Clinic),
Nilay D Shah (Mayo Clinic), Lila J Rutten (Mayo Clinic), Jennifer B McCormick (Mayo Clinic),
Gyorgy Simon (Mayo Clinic)
Other Faculty: Debra Steele-Johnson, Jack Dustin
PhD Students: Ashutosh Jadhav, Amir Hossein Yazdavar, Hussein Al-Olimat
Master Student: Surendra Marupudi
Visiting Scholar: SoonJye Kho
Ohio Center of Excellence in Knowledge-Enabled Computing
NIH Award#: 1 R01 MH105384-01A1
Modeling Social Behavior for Healthcare Utilization in Depression
Timeline: 1 Jul. 2015 - 30 Jun. 2019
Award Amount: $1,934,525 (WSU: $505,600)
Additional Funded Projects (incomplete list)
● PFI: AIR-TT: Market-driven Innovations and Scaling up of Twitris - A System for Collective
Social Intelligence; 200K, Sheth, Mackay
● Medical Information Decision Assistance and Support; 25K, Prasad, Sheth
● Westwood Partnership to Prevent Juvenile Repeat Violent Offenders; $200K, Sheth, Doran,
Dustin
● Semantic Web-based Data Exchange and Interoperability for OEM-Supplier Collaboration;
89K, Prasad, Sheth
● NIDA National Early Warning System Network (iN3): An Innovative Approach; 299K,
Carlson, Sheth, Boyer, Daniulaityte, Nahas
● SemMat: Federated Semantic Services Platform for Materials Science and Engineering;
315K, Sheth, Prasad, Srinivasan
● Materials Database Knowledge Discovery and Data Mining; 190K, Sheth, Prasad, Srinivasan
Unfunded projects:
● Formal RDF graph model-singleton property with applications to provenance, access control;
massively scalable graph querying and storage (PNNL, NLM, IIT-D, W3C, ….).
● Gender-Based Violence (United Nations Population Fund).
Let’s Talk Big & Smart Data @ Kno.e.sis
• Social Media Big Data - Twitris, eDrugTrends
• Sensor/IoT Big Data - CityPulse, kHealth
• Healthcare Big Data - EMR, Prediction
• Biomedical Big Data - SCOONER, Biomarkers from NextGen
Sequence and Proteomics Data
• Big and Smart Data Science Certificate
Kno.e.sis private cloud:
864 cores,18TB RAM, 17TB SSD, 435TB disk
Economic & IP Development
Created Twitris, a Commercial Grade software, which had significant NSF
and AFRL research funds, followed by NSF I-CORP and NSF-PFI-AIR
funding; currently VC/entrepreneurs are evaluating for potential
licensing/start up
One recent patent awarded, two filed
Many more patents filed by companies where Kno.e.sis students intern
ezDI has funded Sponsored Research for the fifth year in the row and now
has major successful products
A local entrepreneur has just signed Sponsored Research that is essentially
incubating his company in Kno.e.sis
On going SBIR/STTR collaborations helping economic development
Examples of Real-world Impact
Natural Disaster/Crisis Response Coordination:
During Hurricane Phallin, Uttarakhand floods, JK floods our
technology was used by digital volunteers for real-time
rescue response coordination, saving lives.
Semantic Web-empowered NLU* of clinical text:
commercialized as part of Computer Assisted Coding, etc.
Application deployed for operational use, open source
ontologies, tools, data, patents, etc.
*Natural Language Understanding
Our Most Important Metric of Success
World-class graduates:
• Innovative, confident, good communicators
• Sound technological & engineering skills
• Well-networked, exceptionally well cited
• Granted top internships and awards, hired to top jobs in their fields, able
to pursue careers of their choosing (academic, industry research, R&D,
entrepreneurship, etc.)
• Exceptionally successful and leaders in their area in a short time
See: http://knoesis.org/amit/students
Even more @
http://knoesis.org
http://knoesis.org/amit
Kno.e.sis:
Ohio Center of Excellence in Knowledge-enabled Computing
top related