Towards Knowledge Graph Profiling
1 like1,880 views
The document discusses the concept and profiling of knowledge graphs, providing insights into their creation, characteristics, and the trade-offs involved in their development. It highlights various popular knowledge graphs, their collection methodologies, and deficiencies, while emphasizing the importance of understanding which graph to use for specific purposes. Additionally, ongoing work to enhance knowledge graph quality and completeness is noted, including the exploration of new knowledge sources beyond Wikipedia.
Towards Knowledge Graph Profiling
- 1. 10/22/17 Heiko Paulheim 1 Towards Knowledge Graph Profiling Heiko Paulheim
- 2. 10/22/17 Heiko Paulheim 2 Introduction • You’ve seen this, haven’t you? Linking Open Data cloud diagram 2017, by Andrejs Abele, John P. McCrae, Paul Buitelaar, Anja Jentzsch and Richard Cyganiak. http://lod-cloud.net/
- 3. 10/22/17 Heiko Paulheim 3 Introduction • Knowledge Graphs on the Web • Everybody talks about them, but what is a Knowledge Graph? – I don’t have a definition either...
- 4. 10/22/17 Heiko Paulheim 4 Introduction • Knowledge Graph definitions • Many people talk about KGs, few give definitions • Working definition: a Knowledge Graph – mainly describes instances and their relations in a graph • Unlike an ontology • Unlike, e.g., WordNet – Defines possible classes and relations in a schema or ontology • Unlike schema-free output of some IE tools – Allows for interlinking arbitrary entities with each other • Unlike a relational database – Covers various domains • Unlike, e.g., Geonames
- 5. 10/22/17 Heiko Paulheim 5 Introduction • Knowledge Graphs out there (not guaranteed to be complete) public private Paulheim: Knowledge graph refinement: A survey of approaches and evaluation methods. Semantic Web 8:3 (2017), pp. 489-508
- 6. 10/22/17 Heiko Paulheim 6 Motivation • In the coming days, you’ll see quite a few works – that use DBpedia for doing X – that use Wikidata for doing Y – ... • If you see them, do you ever ask yourselves: – Why DBpedia and not Wikidata? (or the other way round?)
- 7. 10/22/17 Heiko Paulheim 7 Motivation • Questions: – which knowledge graph should I use for which purpose? – are there significant differences? – would it help to combine them? • For answering those questions, we need knowledge graph profiling – making quantitative statements about knowledge graphs – defining measures – defining setups in which to measure them
- 8. 10/22/17 Heiko Paulheim 8 Outline • How are Knowledge Graphs created? • Key objectives for profiling Knowledge Graphs – Size – Timeliness – Level of detail – Overlap – ... • Key figures for public Knowledge Graphs • Common Shortcomings of Knowledge Graphs – ...and how to address them • New Kids on the Block
- 9. 10/22/17 Heiko Paulheim 9 Knowledge Graph Creation: CyC • The beginning – Encyclopedic collection of knowledge – Started by Douglas Lenat in 1984 – Estimation: 350 person years and 250,000 rules should do the job of collecting the essence of the world’s knowledge • The present – >900 person years – Far from completion – Used to exist until 2017
- 10. 10/22/17 Heiko Paulheim 10 Knowledge Graph Creation • Lesson learned no. 1: – Trading efforts against accuracy Min. efforts Max. accuracy
- 11. 10/22/17 Heiko Paulheim 11 Knowledge Graph Creation: Freebase • The 2000s – Freebase: collaborative editing – Schema not fixed • Present – Acquired by Google in 2010 – Powered first version of Google’s Knowledge Graph – Shut down in 2016 – Partly lives on in Wikidata (see in a minute)
- 12. 10/22/17 Heiko Paulheim 12 Knowledge Graph Creation • Lesson learned no. 2: – Trading formality against number of users Max. user involvement Max. degree of formality
- 13. 10/22/17 Heiko Paulheim 13 Knowledge Graph Creation: Wikidata • The 2010s – Wikidata: launched 2012 – Goal: centralize data from Wikipedia languages – Collaborative – Imports other datasets • Present – One of the largest public knowledge graphs (see later) – Includes rich provenance
- 14. 10/22/17 Heiko Paulheim 14 Knowledge Graph Creation • Lesson learned no. 3: – There is not one truth (but allowing for plurality adds complexity) Max. simplicity Max. support for plurality
- 15. 10/22/17 Heiko Paulheim 15 Knowledge Graph Creation: DBpedia & YAGO • The 2010s – DBpedia: launched 2007 – YAGO: launched 2008 – Extraction from Wikipedia using mappings & heuristics • Present – Two of the most used knowledge graphs
- 16. 10/22/17 Heiko Paulheim 16 Knowledge Graph Creation • Lesson learned no. 4: – Heuristics help increasing coverage (at the cost of accuracy) Max. accuracy Max. coverage
- 17. 10/22/17 Heiko Paulheim 17 Knowledge Graph Creation: NELL • The 2010s – NELL: Never ending language learner – Input: ontology, seed examples, text corpus – Output: facts, text patterns – Large degree of automation, occasional human feedback • Today – Still running – New release every few days
- 18. 10/22/17 Heiko Paulheim 18 Knowledge Graph Creation • Lesson learned no. 5: – Quality cannot be maximized without human intervention Min. human intervention Max. accuracy
- 19. 10/22/17 Heiko Paulheim 19 Summary of Trade Offs • (Manual) effort vs. accuracy and completeness • User involvement (or usability) vs. degree of formality • Simplicity vs. support for plurality and provenance → all those decisions influence the profile of a knowledge graph!
- 20. 10/22/17 Heiko Paulheim 20 Non-Public Knowledge Graphs • Many companies have their own private knowledge graphs – Google: Knowledge Graph, Knowledge Vault – Yahoo!: Knowledge Graph – Microsoft: Satori – Facebook: Entities Graph – Thomson Reuters: permid.org (partly public) • However, we usually know only little about them
- 21. 10/22/17 Heiko Paulheim 21 Comparison of Knowledge Graphs • Release cycles Instant updates: DBpedia live, Freebase Wikidata Days: NELL Months: DBpedia Years: YAGO Cyc • Size and density Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017 Caution!
- 22. 10/22/17 Heiko Paulheim 22 Comparison of Knowledge Graphs • What do they actually contain? • Experiment: pick 25 classes of interest – And find them in respective ontologies • Count instances (coverage) • Determine in and out degree (level of detail)
- 23. 10/22/17 Heiko Paulheim 23 Comparison of Knowledge Graphs Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 24. 10/22/17 Heiko Paulheim 24 Comparison of Knowledge Graphs • Summary findings: – Persons: more in Wikidata (twice as many persons as DBpedia and YAGO) – Countries: more details in Wikidata – Places: most in DBpedia – Organizations: most in YAGO – Events: most in YAGO – Artistic works: • Wikidata contains more movies and albums • YAGO contains more songs Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 25. 10/22/17 Heiko Paulheim 25 Caveats • Reading the diagrams right… • So, Wikidata contains more persons – but less instances of all the interesting subclasses? • There are classes like Actor in Wikidata – but they are hardly used – rather: modeled using profession relation
- 26. 10/22/17 Heiko Paulheim 26 Caveats • Reading the diagrams right… (ctd.) • So, Wikidata contains more data on countries, but less countries? • First: Wikidata only counts current, actual countries – DBpedia and YAGO also count historical countries • “KG1 contains less of X than KG2” can mean – it actually contains less instances of X – it contains equally many or more instances, but they are not typed with X (see later) • Second: we count single facts about countries – Wikidata records some time indexed information, e.g., population – Each point in time contributes a fact
- 27. 10/22/17 Heiko Paulheim 27 Overlap of Knowledge Graphs • How largely do knowledge graphs overlap? • They are interlinked, so we can simply count links – For NELL, we use links to Wikipedia as a proxy DBpedia YAGO Wikidata NELL Open Cyc Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 28. 10/22/17 Heiko Paulheim 28 Overlap of Knowledge Graphs • How largely do knowledge graphs overlap? • They are interlinked, so we can simply count links – For NELL, we use links to Wikipedia as a proxy Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 29. 10/22/17 Heiko Paulheim 29 Overlap of Knowledge Graphs • Links between Knowledge Graphs are incomplete – The Open World Assumption also holds for interlinks • But we can estimate their number • Approach: – find link set automatically with different heuristics – determine precision and recall on existing interlinks – estimate actual number of links Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 30. 10/22/17 Heiko Paulheim 30 Overlap of Knowledge Graphs • Idea: – Given that the link set F is found – And the (unknown) actual link set would be C • Precision P: Fraction of F which is actually correct – i.e., measures how much |F| is over-estimating |C| • Recall R: Fraction of C which is contained in F – i.e., measures how much |F| is under-estimating |C| • From that, we estimate |C|=|F|⋅P⋅ 1 R Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 31. 10/22/17 Heiko Paulheim 31 Overlap of Knowledge Graphs • Mathematical derivation: – Definition of recall: – Definition of precision: • Resolve both to , substitute, and resolve to R= |Fcorrect| |C| P= |Fcorrect| |F| |Fcorrect| |C| Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017 |C|=|F|⋅P⋅ 1 R
- 32. 10/22/17 Heiko Paulheim 32 Overlap of Knowledge Graphs • Experiment: – We use the same 25 classes as before – Measure 1: overlap relative to smaller KG (i.e., potential gain) – Measure 2: overlap relative to explicit links (i.e., importance of improving links) • Link generation with 16 different metrics and thresholds – Intra-class correlation coefficient for |C|: 0.969 – Intra-class correlation coefficient for |F|: 0.646 • Bottom line: – Despite variety in link sets generated, the overlap is estimated reliably – The link generation mechanisms do not need to be overly accurate Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 33. 10/22/17 Heiko Paulheim 33 Overlap of Knowledge Graphs Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 34. 10/22/17 Heiko Paulheim 34 Overlap of Knowledge Graphs • Summary findings: – DBpedia and YAGO cover roughly the same instances (not much surprising) – NELL is the most complementary to the others – Existing interlinks are insufficient for out-of-the-box parallel usage Ringler & Paulheim: One Knowledge Graph to Rule them All? KI 2017
- 35. 10/22/17 Heiko Paulheim 35 Common Shortcomings of Knowledge Graphs • Knowledge Graph Profiling can reveal certain shortcomings – ...and once we know them, we can address them • What is the impact of that? – Example: answering a SPARQL query
- 36. 10/22/17 Heiko Paulheim 36 Finding Information in Knowledge Graphs • Find list of science fiction writers in DBpedia select ?x where {?x a dbo:Writer . ?x dbo:genre dbr:Science_Fiction} order by ?x
- 37. 10/22/17 Heiko Paulheim 37 Finding Information in Knowledge Graphs • Results from DBpedia Arthur C. Clarke? H.G. Wells? Isaac Asimov?
- 38. 10/22/17 Heiko Paulheim 38 Common Shortcomings of Knoweldge Graphs • What reasons can cause incomplete results? • Two possible problems: – The resource at hand is not of type dbo:Writer – The genre relation to dbr:Science_Fiction is missing select ?x where {?x a dbo:Writer . ?x dbo:genre dbr:Science_Fiction} order by ?x
- 39. 10/22/17 Heiko Paulheim 39 Common Shortcomings of Knowledge Graphs • Various works on Knowledge Graph Refinement – Knowledge Graph completion – Error detection • See, e.g., 2017 survey in Semantic Web Journal Paulheim: Knowledge Graph Refinement – A Survey of Approaches and Evaluation Methods. SWJ 8(3), 2017 Tuesday, 4:30 pm Journal track paper presentation
- 40. 10/22/17 Heiko Paulheim 40 New Kids on the Block Subjective age: Measured by the fraction of the audience that understands a reference to your young days’ pop culture...
- 41. 10/22/17 Heiko Paulheim 41 New Kids on the Block • Wikipedia-based Knowledge Graphs will remain an essential building block of Semantic Web applications • But they suffer from... – ...a coverage bias – ...limitations of the creating heuristics
- 42. 10/22/17 Heiko Paulheim 42 Wikipedia’s Coverage Bias • One (but not the only!) possible source of coverage bias – Articles about long-tail entities become deleted
- 43. 10/22/17 Heiko Paulheim 43 Work in Progress: DBkWik • Why stop at Wikipedia? • Wikipedia is based on the MediaWiki software – ...and so are thousands of Wikis – Fandom by Wikia: >385,000 Wikis on special topics – WikiApiary: reports >20,000 installations of MediaWiki on the Web
- 44. 10/22/17 Heiko Paulheim 44 Work in Progress: DBkWik • Back to our original example...
- 45. 10/22/17 Heiko Paulheim 45 Work in Progress: DBkWik • Back to our original example...
- 46. 10/22/17 Heiko Paulheim 46 Work in Progress: DBkWik • The DBpedia Extraction Framework consumes MediaWiki dumps • Experiment – Can we process dumps from arbitrary Wikis with it? – Are the results somewhat meaningful?
- 47. 10/22/17 Heiko Paulheim 47 Work in Progress: DBkWik • Example from Harry Potter Wiki http://dbkwik.webdatacommons.org/
- 48. 10/22/17 Heiko Paulheim 48 Work in Progress: DBkWik • Differences to DBpedia – DBpedia has manually created mappings to an ontology – Wikipedia has one page per subject – Wikipedia has global infobox conventions (more or less) • Challenges – On-the-fly ontology creation – Instance matching – Schema matching
- 49. 10/22/17 Heiko Paulheim 49 Work in Progress: DBkWik Dump Downloader Extraction Framework Interlinking Instance Matcher Schema Matcher MediaWiki Dumps Extracted RDF Internal Linking Instance Matcher Schema Matcher Consolidated Knowledge Graph DBkWik Linked Data Endpoint 1 2 34 5 • Avoiding O(n²) internal linking: – Match to DBpedia first – Use common links to DBpedia as blocking keys for internal matching
- 50. 10/22/17 Heiko Paulheim 50 Work in Progress: DBkWik • Downloaded ~15k Wiki dumps from Fandom – 52.4GB of data, roughly the size of the English Wikipedia • Prototype: extracted data for ~250 Wikis – 4.3M instances, ~750k linked to DBpedia – 7k classes, ~1k linked to DBpedia – 43k properties, ~20k linked to DBpedia – ...including duplicates! • Link quality – Good for classes, OK for properties (F1 of .957 and .852) – Needs improvement for instances (F1 of .641) Monday 6:30pm Poster presentation
- 51. 10/22/17 Heiko Paulheim 51 Work in Progress: WebIsALOD • Background: Web table interpretation • Most approaches need typing information – DBpedia etc. have too little coverage on the long tail – Wanted: extensive type database
- 52. 10/22/17 Heiko Paulheim 52 Work in Progress: WebIsALOD • Extraction of type information using Hearst-like patterns, e.g., – T, such as X – X, Y, and other T • Text corpus: common crawl – ~2 TB crawled web pages – Fast implementation: regex over text – “Expensive” operations only applied once regex has fired • Resulting database – 400M hypernymy relations Seitner et al.: A large DataBase of hypernymy relations extracted from the Web. LREC 2016
- 53. 10/22/17 Heiko Paulheim 53 Work in Progress: WebIsALOD • Back to our original example... http://webisa.webdatacommons.org/
- 54. 10/22/17 Heiko Paulheim 54 Work in Progress: WebIsALOD • Initial effort: transformation to a LOD dataset – including rich provenance information Hertling & Paulheim: WebIsALOD: Providing Hypernymy Relations extracted from the Web as Linked Open Data. ISWC 2017
- 55. 10/22/17 Heiko Paulheim 55 Work in Progress: WebIsALOD • Estimated contents breakdown Hertling & Paulheim: WebIsALOD: Providing Hypernymy Relations extracted from the Web as Linked Open Data. ISWC 2017
- 56. 10/22/17 Heiko Paulheim 56 Work in Progress: WebIsALOD • Main challenge – Original dataset is quite noisy (<10% correct statements) – Recap: coverage vs. accuracy – Simple thresholding removes too much knowledge • Approach – Train RandomForest model for predicting correct vs. wrong statements – Using all the provenance information we have – Use model to compute confidence scores Hertling & Paulheim: WebIsALOD: Providing Hypernymy Relations extracted from the Web as Linked Open Data. ISWC 2017
- 57. 10/22/17 Heiko Paulheim 57 Work in Progress: WebIsALOD • Current challenges and works in progress – Distinguishing instances and classes • i.e.: subclass vs. instance of relations – Splitting instances • Bauhaus is a goth band • Bauhaus is a German school – Knowledge extraction from pre and post modifiers • Bauhaus is a goth band → genre(Bauhaus, Goth) • Bauhaus is a German school → location(Bauhaus, Germany) Hertling & Paulheim: WebIsALOD: Providing Hypernymy Relations extracted from the Web as Linked Open Data. ISWC 2017 Tuesday, 2:30 pm Resource track paper presentation
- 58. 10/22/17 Heiko Paulheim 58 Take Aways • Knowledge Graphs contain a massive amount of information – Various trade offs in their creation – That lead to different profiles – ...and different shortcomings • Knowledge Graph Profiling – What is in a knowledge graph? – At which level of detail is it described? – How different are knowledge graphs? • Various methods exist for – ...addressing the various shortcomings • New kids on the block – DBkWik and WebIsALOD – Focus on long tail entities
- 59. 10/22/17 Heiko Paulheim 59 Towards Knowledge Graph Profiling Heiko Paulheim