際際滷

際際滷Share a Scribd company logo
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Related-Work.net
a scientific discussion platform
WeST Koblenz
21.2.2012
Heinrich Hartmann
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Plan
 Academic knowledge discovery
 Vision of Related-Work.net
 System details and open problems
 Demo
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Academic knowledge discovery
 Finding and filtering
publications
 Connect people
interested in the same
paper
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Solution: The Academic Graph
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Problem: No Open Access!
No Open Access:
* Citation data
* Full-text
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Problem: No Open Access!
No Open Access:
* Citation data
* Full-text
* Social information needs
to be provided by
community
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Existing services have shortcomings
Citation Data Community Open Source/Data
Google Scholar Yes No No/No
Microsoft Academic Yes No No/No
SciVerse (Elsevier) Yes No No/No
Mendeley Yes Yes No/No
ResearchGate Yes Yes No/No
CiteSeerX Yes (quality?) No Yes/Yes (broken)
dblp No No -/Yes
Bibsonomy No Yes Yes/Yes
Related-Work.net Yes Yes Yes
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Plan
 Academic knowledge discovery
 Vision of Related-Work.net
 System details and open problems
 Demo
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Vision for Related-Work.net
 Social community for scientists
 Open database of papers and citations
 Free software
 Strong data mining:
Recommender, Auto completion, News feed
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Vision of Related-Work.net
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
History
 March '12
 Idea Heinrich & Rene: write proposal
 Sept '12
 Heinrich quit Maths
 Writes Citation extraction for Arxiv.org
 Networking in Oxford (Akorn, OpenCitations)
 Dec '12
 Merger with OpenCitationsCorpus by David Shotton
 JISC Grant: Cottage Labs
 GWTP protoype development
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Team: Related-Work.net / OpenCitations.net
Ren辿
Pickhardt
Mathematics /
Computer Science
co-founder RW.net
metalcon.de
Heinrich
Hartmann
Mathematics /
Computer Science
co-founder RW.net
David
Shotton
Oxford Zoologist
OpenCitations.net
CiTO/SPAR Ontologies
JISC Grant:
Cottage Labs
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Open Citations and Semantic Publishing
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Cottage Labs
Richard
Jones
Mark
MacGillivray
Martyn
Whitewell
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Plan
 Academic knowledge discovery
 Vision of Related-Work.net
 System details and open problems
 Demo
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Data Ingest Pipeline
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Example Matching Problem
 A.G. Bashkirov, Physica A 340 , 153 (2000)
 very little information
 Robert L. Pego and Michael I. Weinstein. Eigenvalues, and instabilities of solitary
waves. Philos. Trans. Roy. Soc. London Ser. A , 340(1656):47--94, 1992.
 Not Arxiv
 D. V. Shirkov and I. L. Solovtsov, Theor. Math. Phys. 150 , 132 (2007) arXiv:hep-
ph/0611229 .
 found ID: hep-ph/0611229
 G.J. Galloway, Maximum principles for null hypersurfaces and null splitting theorems ,
Journal APPT 1 543-567 2000 .
 year author title heuristic: math/9909158
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Example Matching Problem
 A.G. Bashkirov, Physica A 340 , 153 (2000)
 very little information
 Robert L. Pego and Michael I. Weinstein. Eigenvalues, and instabilities of solitary
waves. Philos. Trans. Roy. Soc. London Ser. A , 340(1656):47--94, 1992.
 Not Arxiv
 D. V. Shirkov and I. L. Solovtsov, Theor. Math. Phys. 150 , 132 (2007) arXiv:hep-
ph/0611229 .
 found ID: hep-ph/0611229
 G.J. Galloway, Maximum principles for null hypersurfaces and null splitting theorems ,
Journal APPT 1 543-567 2000 .
 year author title heuristic: math/9909158
Extracted
16 Mio. citation strings
only 2 Mio.
currently matched!
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Author Identification Problem
 Two authors w. same name
 Author changes name
Approaches:
 Official author IDs (ORCID / Arxiv / PMC )
 Graph Mining
 Email addresses (!)
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Email addresses are in the FullText!
Low Redshift QSO Lyman alpha Absorption Line Systems
Associated with Galaxies
 W.P. Lin
 G. Boerner
 H.J. Mo
 linwp@bac.pku.edu.c
 hom@mpa-garching.mpg.de
 linwp@mpa-garching.mpg.de
 grb@mpa-garching.mpg.de
Found
750.000
Email addresses
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Standard Architecture for the Front End
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Data Mining Examples 1
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Data Mining Example 2
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Data Mining Example 3
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Open Problems
 Crowdsourcing citation data
Improve matching algorithms (by ML? ActiveLearning?)
 Community building / policy modeling
 Add further data sources (cur. ArXiv/PMC/CrossRef)
 Automated schema detection
Heinrich Hartmann
hartmann@uni-koblenz.de
WeST Institute
Plan
 Academic knowledge discovery
 Vision of Related-Work.net
 System details and open problems
 Demo

More Related Content

Related-Work.net at WeST Oberseminar