1. "How semantic technology enhances the productivity of
scientific researchers."
Darrell W. Gunter
Collexis an Elsevier Company
August 5, 2010
2. Our Agenda For Today
The Challenges of Scientific Research
The Collexis Technology
Case Studies
Professional Networks
Institutional Networks Johns Hopkins & Asklepios
Managing the Peer Review Process
Summation
3. The Facts
• 90 – 100 hours to
write an article
Researchers • 2 – 3 Peer Review 3 –
6 hrs
• Articles 154 vs. 83
• # pages/article 12.4 vs.
Journal Growth 7.4
• Total pages 2,216 vs.
820
23,000 Jnls / 90%
electronic /Articles –
800K+
The Author is under great pressure!
1,2,3
5. Collexis Knowledge Engine 7.0
Abbreviation
Tokenizer Normalizer
expansion
Language Coordination
Dehyphenation
detection expansion
Part-of-Speech Entity recognition Noun phrase
tagging based on regular expressions detection
Part-of-speech Exclude known
Concept finding based disambiguation of
thesaurus concepts idioms
Fingerprint
aggregation
6. Collexis Knowledge Engine 7.0
Modular NLP workbench – processing and analyzing of text
documents
Retrieval and aggregation engine – serving the application
layer
7. Collexis – selected references
Dana Farber Cancer Institute Asklepios Kliniken Johnson & Johnson
Harvard University
National Institutes of Health Johns Hopkins University University of California, San
Franciscio
American Institute of Physics Mayo Clinic Stanford University
The Wellcome Trust California Institute for Albert Einstein College of
Quantitative Biosciences (QB3), Medicine
9. Creating expert profiles from documents using
semantic technologies
Document fingerprints aggregated to expert profiles!
10. BiomedExperts – more than 300,000+ registered users
Prepopulated network – based on PubMed
1.8 million precalculated experts
More than 24 million co-author relations between them
Representing over 3,500 institutions
From 190 countries
Growing each day between 500 and 1000 users
BME data used in other applications
16. Johns Hopkins: The Issue! Connecting Experts
Fall retreat: Main issue how do they take advantage of the
university‟s expertise and build collaboration
First solution – Repurpose a parking lot to be a coffee shop
for the JH community to grab a cup of Joe and find new
collaborators.
Outcome – Great coffee, great conversation but
collaboration did not take off.
The Collexis Solution – Expert Institutional Dashboard!
22. Asklepios
Facts and Figures
• Asklepios - Europe„s largest health care
provider
– 500.000 patients for inpatient care per year, 95 hospitals, 21.000 beds
– 34.500 employees
– Asklepios owns medical nursing and allied health schools
– Home care programs and residential care programs
• Asklepios International
– Pacific Health System – California
– Greece, Athens Medical Center
– University hospital in Shanghai: Joint Venture with Siemens and Tongji University
23. Why Knowledge
Management?
Guide Workflows Optimize Workflows Distribute
(e.g. Care Plans, Expert- (e.g. Avoid interruptions Expert Knowledge
Task Context Allocation) caused by knowledge search, (across multiple locations,
retrieval, and application) time zones, medical
conditions)
Stimulate new
Help Asklepios to know
Knowledge Acquisition Usage
“what Asklepios knows”
Models
24. Use Case 1 – Expert
profiles
• Patient, male, age of 62, needs a knee joint prosthesis
due to Rheumatoid Arthritis
• Where is the best place to get it?
• Criteria which will be taken into account:
– Geographical aspects
– Recommendation of his GP
– Publicly available information - mostly via Internet
• Strongest competitors: university hospitals (within the
region)
28. Provide a single point of search
for all relevant content from
publishers!
Link internal expertise /
experience and external
knowledge sources!
29. Use Case 4 - External Resources and
Internal Experience Use Case 4 - External
Resources and Internal Experience
• Patient with lung cancer and reduced renal function
• Decision in chemotherapeutic drug is pending
• Preferred choice: Cisplatin as chemotherapeutic agent
• Open questions: can Cisplatin be used which has
nephrotoxicity as a side effect?
31. Link External Knowledge and
Internal Expertise
Opening an journal article…
…shows immediately similar
publications colleagues
… and the names
and expert profiles
32. Key Issues in STM Industry
Publishers / Editors
Finding the right reviewer
Expanding their pool of reviewers
Institutions
Determining what grants they should go after
Determining who within their organization is best to apply
Grant Funding Organizations
Analyzing the vast amount of grant applications submitted.
Determining who within the organization is best qualified to review the grant application
(known and unknown)
33. The Challenge for STM Publishers
Receive thousands of manuscripts annually
Timely process to conduct the Peer Review Process
Timely process to determine who should review it.
Important for reviewer to free of conflicts of interest
Ethics of review process are paramount
34.
35.
36.
37.
38. Key Benefits Reviewer Finder
Fingerprint of manuscript - Clarity
Determine the best reviewer
Free of conflicts
More efficient and effective process
Ultimately increases profitability
39. The effectiveness of Semantic Technology
Aggregates the researcher's publications into a Fingerprint of weighted
relevant concepts
Expert Profiles (individual, institution, dept, country,etc.)
Shows co-author relationships (who publishes with whom)
Conduct search by key concepts
Match content from a variety of sources based on a key concept,
researcher, country, etc.
Determine expert for peer review, grant application, project, etc.
40. Thank you for your attention!
www.collexis.com
Darrell W. Gunter
gunter@collexis.com, cell +1-973-454-3475