Ohio Center of Excellence in Knowledge-Enabled Computing at Wright State (Kno.e.sis)
Center overview: http://bit.ly/coe-k
Invitation: http://bit.ly/COE-invite
13. Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis) scientiapotentiaest Knowledge is Power Francis Bacon, 1597 …established and popularized deductive methodologies for scientific inquiry 3
14. Ohio Center of Excellence Knowledge-Enabled Computing (Kno.e.sis) Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis) Our society has progressed from Agriculture Industrial Service Knowledge … we live in a Knowledge Society 4
15. Knowledge Economy As the economies have transformed …. … new era … where the principal component of value creation, productivity and economic growth is knowledge. Florida & Kenny 91 5
16. Knowledge Economy Thousands knowledge-intensive services are leading the all sector in job creation, R&D spending, average wages and growth. 6
17. Knowledge Economy Exploitation of Knowledge for improving human experience/effectiveness, social and economic gains Computer + Human Knowledge (knowledge discovery, insight) Computer + Collective Action Collective Intelligence (Knowledge) Computer (IR,ML,NLP) + Knowledge = (better extraction, next gen search) learning, insight Humanknowledge + Computer Representation = Situated Cognition Computer + Human + Knowledge(better) Interaction, (improved) Human Experience, (higher) Productivity 7
18. Knowledge Services Creation and exploitation of Knowledge for improving human experience/effectiveness, social and economic gains Computer + Human Knowledge (knowledge discovery, insight) Computer + Collective Action Collective Intelligence Computer (IR,ML,NLP) + Knowledge = (better) learning (extraction, search…) Humanknowledge+ Computer Representation = Situated Cognition Computer + Human + Knowledge (better) Interaction, (improved) Human Experience, (higher) Productivity 8
19. affects Migraine Magnesium Stress isa inhibit Knowledge discovery Patient Calcium Channel Blockers 2D-3D & Immersive Visualization, Human Computer Interfaces Impacting bottom line Domain Models/Knowledge Structured text (Scientific publications / white papers) Biomedical Knowledge Discovery, Knowledge Management & Visualization SEMANTICS, MEANING PROCESSING Patterns / Inference / Reasoning Meta data / Semantic Annotations Search and browsing Metadata Extraction/Semantic Annotations Massive amounts of data Clinical Trial Data Experimental Results Public domain knowledge (PubMed) 9
20. Kno.e.sis’ leadership in semantic processing will contribute to basic theory about computation and cognitive systems, and address pressing practical problems associated with productive thinking in the face of an explosion of data. Kno.e.sis intends to lead a march from information age to meaning age. Kno.e.sis Vision 10
43. Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis) Knowledge & Next Generation of the Web Web has become the core infrastructure for the knowledge economy Web 1.0: Web of Documents and Media Web 2.0: Web of People Web 3.0: Web of Meaning MeenaNagarajan Advanced Data Management
50. The Memex Vision [Vannevar Bush] Of now the human brain navigates an information space … “It operates by association. With one item in its grasp, it snaps instantly to the next that is suggested by the association of thoughts, in accordance with some intricate web of trails carried by the cells of the brain.” 24
57. Ohio Center of Excellence Knowledge-Enabled Computing (Kno.e.sis) SEMANTIC SOCIAL WEB
58. Everyone Wants to talk …and be heard! Hundreds and thousands of tweets, facebook posts, blogs about a single event, multiple narratives, strong opinions, breaking news.. 29
59. TWITRIS : Twitter+Tetris Our attempt to help you keep up with citizen observations on Twitter WHAT are people saying, WHEN, from WHERE Puts citizen reports in context for you by overlaying it with news, wikipedia articles! 30
61. Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis) Biomedical & Health Sciences From data to understanding Mike Raymer Many biomedical collaborations: Nick Reo, Toxicology; Tim Cope, Neuroscience; Jerry Alter, Protein Science; Oleg Paliey, Microbiolgy and health Many health care collaborations: Kate Cauley, Center for Healthy Communities; UC-Irvine – Emergency health, Sonia Michail, obesity & intestinal health; Bradley Jacobs, human health Human Sciences & Health Care
80. Nationwide Health CDC VA IHS DoD SSA State and Local Gov Health Bank orPHR Support Organization Community Health Centers Community #1 Labs As part of Nationwide Health Information Network effort, Center for Healthy Communities (CHC) will use its HIExTM system, supported through Wright State HealthLink and medical providers to electronically transmit data from their certified EHRs to the Social Security Administration (SSA). This will improved quality of care through analysis of large data sets documenting treatments and outcomes. Knoesis will be in the forefront creating the systems intelligence to better understand this complex, integrated information both at the individual provider level and in the realm of population health. CHC received ~1M contact from SSA. IntegratedDelivery System Pharmacies Community #2 38
81. Biomedical & Health Sciences From data to understanding Data Information Understanding 39
82. Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis) Cognitive Science & applications to Human Effectiveness John Flach Defense/Aerospace R & D
83. Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis)
92. Regional Impact Collaboration by daytaOhio and the Dayton Development Coalition to Leverage Kno.e.sis Terry Rapoch – President/CEO daytaOhio Jim Leftwich – CEO Dayton Development Coalition 50
93. Regional Development Model Defense/Aerospace R & D Knoesis Advanced Data Management Human Sciences & Health Care 51
115. Board of Advisor’s View Prof. Ahmed Elmagarmid Cyber Center – Discovery Park Purdue University Ohio Center of Excellence on Knowledge-Enabled Computing (Kno.e.sis)
116.
117. Prof. Sheth among the most cited Computer Science authors in the world today(top 30 based on h-index)
118. Prof. Bennett & Flach’s paper declared as one of most influential papers published in over 50 years in Journal of Human Factors; Prof. Raymer’s paper was cited in a US Supreme Court decision
123. Funding Starting with current active funds of $8-10 million (supporting research of 15 faculty and 45+ funded grad students & postdocs) Kno.e.sis anticipates growing to $13 million in 5 years and $19.5 million in 10 years (25 faculty plus 75 researchers) 59
128. Students interned at & collaborated with the very best places: Microsoft Research, Yahoo! Research, IBM Research, HP Labs, NLM, …and filed for 6 patents in 2 years60
143. PanelWhat are exciting and emerging areas for careers? And how to prepare yourself for one? Prof. Ahmed Elmagarmid, Purdue University Dr. Daniel Gruhl, IBM, Research Dr. PankajMehra, HP Labs Dr. Daniel Serfati, Aptima Dr. Harry Silver, LexisNexis
Let me start by offering my appreciation for our Chancellor Dr. Fingerhut’s visionary leadership in establishing the Ohio Center of Excellence program that identifies Centers and program that generate world-class research and help draw talent and investment to the state. I would be remiss if I did not call out tremendous leadership that our President Dr. Hopkins and his entire leadership team has shown in regards to identifying and promoting these centers– my tanks to Dr. Angle, Dr. Bantle. Dr. Jang, and Dr. Sudkamp– thanks for your early and steadfast support for Kno.e.sis.
In the 16th Century, Francis Bacon—an Englishphilosopher, statesman, scientist, lawyer, jurist and author established deductive methodologies for scientific enquiry… and he said something that is more true than ever– Knowledge is Power
A study of OECD countries over a decade showed that as much as 75% of the economic growth and most new jobs was due to knowledge-intensive services- and the importance of these services was not only limited only high-tech or IT areas, but pervaded all sectors of economy
Let me give a technological introduction to what our center is about: we all face a fire hose of data-- Pubmed adds 2000 to 4000 citations per day, it is usual to add about 5 gig from a single run of a scientific experiment -- and just imagine how much data created by all the cameras and 40 billion mobile sensors in the world! But even with all the search and browsing tools we have, we face huge information glut. How do we make sense from the data? Just as humans apply their knowledge and experience to understand what they see– we apply domain model or knowledge to attach meaningful labels to these data. Then we can apply computational techniques to visualize, provide situational awareness, discovery nuggets of knowledge of information and insight. For example, from all that biomedical data, what a scientist may be looking for is– how can we treat Migraine? What has Magnesium to do with Migraine? Why does Magnesium deficiency cause Migraine? What is the process by which Magnesium affects Migraine?
So what is Kno.e.sis about– it is about stepping away from the concerns of storing and searching data, to that of improving human experience, human effectiveness, human performance, human productivity.
Our 15 faculty from 4 colleges are already engaged in multiple jointly funded grants, pending proposals, serving on interdisciplinary programs like Biomedical Sciences PhD program and on committees of students of colleagues.
One of the hallmarks of the knowledge enabled computinginfrasturucture is Enabling machines to understand the data that they are processing. At the core of this vision also shared by the Semantic Web is having a formal representation of a domain or what we call ‘domain knowledge’ –the entities involved in a domain and how are they related to each other..Where do we get some domain knowledge from – one could look to domain experts to categorize world knowledge in that domain, borrow from legacy systems or in many cases tap into the knowledge that is already present in documents ..
One of our projects in this space looks into scalable harvesting of community knowledge that exists about a domain in collaborative sources such as Wikipedia and scientific corpus to construct formal models of a domain
Here is an example of a Human performance n cognition Ontology developed by us using this technique and which is being used by human effectiveness directorate at the AFRLSeveral part of this work were conducted in collaboration with Hewlett Packard research labs who currently conduct beta testing of a commercial version called Taxonom.com
One application of building domain models is knowledge discovery or deriving informational nuggets present in data sources by using facts that you already know about a domain
One of our efforts in this direction uses existing domain models such as the UMLS or those built from wikipedia to guide browsing of concepts in documents and use the browse trails to formulate hypotheses to discover knowledge that we did not know previously existed in the data.
This work that envisions our information space as a web of relationships to connect diferent document or web resources represents in many ways the beginnings of the realization of the expansive memex vision and trailblazing that Dr Vanevar Bush outlined in 1945… As he was speaking to how our human brain navigates the information space… using an association of thoughts that link one item to the next
Several tools that we built in this space are available online and also shared as an open source project with the community
A second area where we extensively employ domain knowledge is in interpreting various forms of sensor data in order to achieve meaningful situational awareness..Today there are more than 40 billion mobile sensors on the groundAnd nearly 4 million people carrying mobile phones taking pictruesWith so many sensors around there is a lot of information available at hand surrounding any eventHow do u know that they all relate to the same event and lend an understanding for how event evolves?
The knowledge enabled techniques we develop enable a fusion of Sensor data – lending understanding from the stages of identifying sensors, to gathering processing and using this data for situational awareness and sense makingSome of our work in this space is being supported by the sensor directorate at the AFRL next door.
The last representative work we’d like to share with you is our work on making sense of social data, like those from Twitter and facebookaround news worthy events that are of interest to a populace.The goal is to offer an understanding of what people are talking about and paying attention to
What the social perceptions behind the data might be, the multiple narratives
Twitris is our effort in this direction to help users keep up with observations made around news-worthy events.. Before I hand over the microphone to Dr. Mike Raymer, I’d like to leave you with a short demo of the deployed web application.
No where is it more important to enable the path from data to understanding than in the medical and life sciences!The barriers to interdisciplinary collaboration are high – natural scientists and computer scientists speak different languages.More new words in their first biology class than an entire year of a foreign language.We have demonstrated success in breaking down those barriers.Cross-disciplinary scientists (we make it our job) & cross disciplinary students.
No where is it more important to enable the path from data to understanding than in the medical and life sciences!The barriers to interdisciplinary collaboration are high – natural scientists and computer scientists speak different languages.More new words in their first biology class than an entire year of a foreign language.We have demonstrated success in breaking down those barriers.Cross-disciplinary scientists (we make it our job) & cross disciplinary students.
No where is it more important to enable the path from data to understanding than in the medical and life sciences!The barriers to interdisciplinary collaboration are high – natural scientists and computer scientists speak different languages.More new words in their first biology class than an entire year of a foreign language.We have demonstrated success in breaking down those barriers.Cross-disciplinary scientists (we make it our job) & cross disciplinary students.
Example – AFRL problem – collaboration with Dr. Nick Reo, School of MedicineDr. Reo – Nuclear Magnetic ResonanceProblem – highly complex data sets – many noisy results from each experiment
Example – AFRL problem – collaboration with Dr. Nick Reo, School of MedicineDr. Reo – Nuclear Magnetic ResonanceProblem – highly complex data sets – many noisy results from each experiment
No where is it more important to enable the path from data to understanding than in the medical and life sciences!The barriers to interdisciplinary collaboration are high – natural scientists and computer scientists speak different languages.More new words in their first biology class than an entire year of a foreign language.We have demonstrated success in breaking down those barriers.Cross-disciplinary scientists (we make it our job) & cross disciplinary students.