A survey of top-level ontologies: framework and results University of Westminster – 20 th January 2021 Chris Partridge, Andrew Mitchell, BORO Solutions, University of Westminster structure Background context term-ology TLO survey & ontological assessment TLO survey candidate top-level ontologies assessment framework Ontological assessment framework: industry data models Summary 2 Background: context organisational 3 origins Launched in July 2018, the National Digital Twin programme was set up to deliver key recommendations of the National Infrastructure Commission 2017 “Data for the Public Good Report” to steer the successful development and adoption of the Information Management Framework for the built environment to create an ecosystem of connected digital twins – a national digital twin– which opens the opportunity to release value for society, the economy, business and the environment deep organizational structure Digital Framework Task Group (DFTG) National Digital Twin programme (NDTp) www.cdbb.cam.ac.uk/what-we-do/national-digital-twin-programme www.cdbb.cam.ac.uk www.gov.uk/government/organisations/department-for-business-energy-and-industrial-strategy www.constructioninnovationhub.org.uk Construction Innovation Hub (CIH) Department for Business, Energy & Industrial Strategy (BEIS) Centre for Digital Britain (CDBB) 5 key milestones The 2018 publication of the Gemini Principles, a paper setting out the proposed principles to guide the national digital twin and the information management framework that will enable it The 2018 publication the Roadmap, a prioritised plan for five core streams responsible for the delivery of the information management framework The 2020 publication of the Pathway towards an Information Management Framework: a 'Commons' for a digital built Britain and a high-level summary The approach to delivering a National Digital Twin for the United Kingdom the Information Management Framework UK’s National Digital Twin programme ( NDTp ) Information Management Framework (IMF) www.cdbb.cam.ac.uk/news/pathway-towards-IMF a national system for connecting digital assets designed to enable competition on delivery and encourage innovation and development over time style.visibility style.visibility style.visibility style.visibility the NDTp has deep levels of components Reference Data Library (RDL) Integration Architecture (IA) Foundation Data Model (FDM) 8 Information Management Framework (IMF) The NDTp has several components, one of which is the IMF A high-level definition of the structure and meaning of data to enable the consistent sharing of data across Digital Twins and the ecosystems they support top-level ontology – in context 9 This TLO will need to be a foundation for data such as UNICLASS The FDM has several components, one of which is the TLO The TLO is a key to the success of the FDM 10 Background: term-ology The -ology (and – metry ) terms we need -ology: a subject of study; a branch of knowledge - metry : the process of measuring two senses of ontology Foundational Ontology ontology is: “the set of things whose existence is acknowledged by a particular theory or system of thought.” [1] [1]. E. J. Lowe in the Oxford Companion to Philosophy. Semantic Web Ontology ontology is: an “explicit specification of a conceptualization” [1] a “formal specification of a shared conceptualization” [2] “a formal, explicit specification of a shared conceptualization.” [3] [1]. T. R. Gruber. A Translation Approach to Portable Ontologies. Knowledge Acquisition, 5(2):199–220, 1993. [2]. W. Borst. Construction of Engineering Ontologies. PhD thesis, Institute for Telematica and Information Technology, University of Twente, Enschede, The Netherlands, 1997 [3] R. Studer, R. Benjamins, and D. Fensel . Knowledge engineering: Principles and methods. Data & Knowledge Engineering, 25(1–2):161–198, 1998 more terms – epistemology and agentology {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} ontology what is god-like view – view from nowhere epistemology what is known view from somewhere agentology what is known and done (by the agent-system) view from an agent Approach: Ontology is the foundation Epistemology/Agentology is expressed in terms of the ontology Agentology is an expansion of epistemology It arises from the recognition that one cannot ontologise away all the epistemology for more see: https://www.academia.edu/35739957/ even more terms – mereology and onomatology {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} mereology wholes and parts the study of parts and the wholes they form onomatology names the study of the etymology, history, and use of proper names Approach: Good examples of obvious ‘universal’ patterns Passes the Groucho Marx four-year-old child test: Rufus T. Firefly: Clear? Huh! Why a four-year-old child could understand this report! Run out and find me a four-year-old child, I can't make head or tail of it Interesting because: clear, obvious, general patterns, also usually are obviously missing from computer information system and even more terms – mereotopology, morphology and geometry {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} mereotopology mereology plus topology adds connections between parts morphology shapes form and structure – includes homology (hole-ology) geometry measurement distance, size, and relative position of figures. Approach: Mereology is a good base, but it needs extending with topology, morphology and then geometry Use a component-based approach to build from the base Background: TLO survey & ontological assessment a survey of top-level ontologies: a framework To inform the ontological choices for a Foundation Data Model Appendix E: Summary of Framework Assessment Matrix Results 31 ontological choices 37 top ontologies shortlisted and assessed The ontological choices shape the architecture of the ontology making an ontological assessment Evidence-based Conceptual Prototyping Mining the ontological requirements for a domain For an example based upon UNICLASS see: https://www.academia.edu/44210409 https://www.academia.edu/44326217 or https://borosolutions.net/multi-level-types-uniclass-multi-2020 Ontological Framework Assessment Two useful tools – among many Both can be applied to any data schema (or data) TLO survey: in detail TLO survey – in context 20 style.visibility TLO survey approach The approach has three parts collect candidate top-level ontologies develop assessment framework assess candidate top-level ontologies against the framework 21 TLO survey deliverable - report initiated by BORO and reviewed by the FDM working group currently in the editing process components main document additional tables candidate top-level ontologies list TLOs – framework assessment matrix 22 style.visibility style.visibility style.visibility style.visibility TLO survey: candidate top-level ontologies 23 top-level ontology candidates a long list of around forty possible candidates for top-level ontology (TLO) content was compiled the net was thrown as wide as possible to identify as much useful content as possible thus, though the focus is on ontological commitment, the list includes data models that are generic in nature (ones without an explicit ontological foundation) as these are likely to have some useful ontological content we anticipate this list will be updated as new TLOs are found (there are around ten additional candidates waiting to be assessed) as it currently stands, it provides a reasonably comprehensive picture of what is currently available 24 initial findings diversity there is an incredible diversity of top-level structure among the candidates some local commonalities, but prima facie – no clear global organising pattern aspiration for universality is common architects commonly build in a universal top-level in other words, this is a common architectural approach/pattern this is particularly noticeable when their domain is quite specialised range of levels of explicit ontological commitment from none to high 25 examples – in no particular order Dublin Core - DCMI resource model MarineTLO: A Top-Level Ontology for the Marine Domain SUMO WordNet OWL – Web Ontology Language 26 style.visibility style.visibility style.visibility style.visibility style.visibility style.visibility TLO survey: assessment framework 27 a framework for choosing ontological commitments an ontology is (according to Jonathon Lowe in The Oxford Companion to Philosophy) “the set of things whose existence is acknowledged by a particular theory or system of thought.” when we interpret a dataset, working out what the data refers to we are acknowledging that the dataset commits to these things existing it turns out that there are a variety of ways of making (choosing) these commitments a top-level ontology enables one to make the choice of ontological commitments in an explicit and consistent way there have been a series of attempts to get to grips with the kinds of choices of ontological commitments that ontologies can make these provide a reasonable starting point, but need substantial further work to provide a comprehensive framework for making choices across a broad range of commitments we had to develop a comprehensive framework 28 assessment framework development driven by two broad considerations general ontological requirements the requirement for an overarching ontological architecture 29 our developed framework’s structure a framework for surveying the ontological commitments {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} level name description 1 general the general approach to ontological commitment 2 formal structure characteristics of the overall formal structure that arises as a result of the individual ontological commitments 3 universal the universal-level individual core commitments 30 general choices TLO survey: assessment framework: framework level 1 31 the general choices {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} type choice description ontologically committed ontological or generic generic focuses on the data structures and makes no explicit ontological commitments categorical yes or no whether the categories are intended to be comprehensive {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} type choice description commitment level high or low the broad level of ontological commitments subject foundational or natural language natural language ontology - what a community implicitly accepts when using a language foundational ontology - what ‘really’ exists according to science (and philosophy) if ontological commitments are made explicitly; general choices characterise a TLO’s broad approach to ontological commitments 32 style.visibility style.visibility style.visibility style.visibility example – generic versus ontological generic ontological UFO GFO 33 style.visibility style.visibility style.visibility style.visibility example – foundational versus natural language foundational natural language WordNet BFO 34 style.visibility style.visibility style.visibility style.visibility example – low versus high commitment low high MarineTLO 35 style.visibility style.visibility style.visibility style.visibility framework assessment matrix: general choices 36 style.visibility style.visibility general choices – visualised note outlier 37 style.visibility general choices – percentages 38 {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} ontologically committed generic 32.4% ontological 59.5% {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} commitment level high 45.5% low 54.5% overall formal structure TLO survey: assessment framework: framework level 2 39 overall formal structure {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} adopted term alternative terms conventional upwards direction examples whole-part part of part-to-whole My arm has a whole-part relation to my hand (my hand is part of my body) type-instance instantiation, class-member, member, instance instance-to-type Arm has a type-instance relation to my arm (my arm is an instance/member of the type arm – my arm is an arm) super-sub-type generalisation, subsumption, super-type, sub-type subtype-to-supertype Limb has a super-sub-type relation to arm and legs (arm and leg are sub-types of limb – arms and legs are limbs) the formal structure reveals itself in three core hierarchical relations 40 basis TLO survey: assessment framework: 41 simplicity simplicity provides us with a good basis for a broad assessment simplicity can be thought of as having two aspects structural where structural or syntactic simplicity is roughly concerned with the shape of the organising structure reflected in the overall formal structure of the TLO and ontological ontological simplicity is roughly concerned with the number of objects 42 ontological simplicity: plenitude versus promiscuity ontological simplicity accounting a variant of Occam's razor fundamental entities cost more than derived entities so minimise fundamental entities a basis for assessing the TLO’s choices of commitment plenitude fundamental entities that lead to many useful derived entities what emerges approach is a general pressure towards a permissive and abundant view of what there is, coupled with a restrictive and sparse view of what is fundamental standard examples: set theory and classical mereology typical hallmark: fruitfulness promiscuity (profligacy, overgeneration, useless ‘plenitude’) fundamental entities that lead to too many useless derived entities typical hallmark: lack of fruitfulness 43 overall formal structure: simplicity aspects the ontological choices leave their mark on this formal structure, the ontological architecture they do this in two characteristic ways which we tag vertical and horizontal aspects these roughly mirror the two types of simplicity discussed above {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} vertical aspect structural or syntactic simplicity deals with the structure of the hierarchies; roughly their shape up and down horizontal aspect ontological simplicity deals with the broad ontological choices that can introduce a division across the hierarchy (horizontal stratification) 44 vertical aspect: varieties of hierarchies TLO survey: assessment framework: overall formal structure - framework level 2 0.2 45 vertical aspect choices {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} type type description relation characteristic choice parent-arity the number of parents allowed type-instance single or unconstrained super-sub-type single or unconstrained transitivity whether aRb and bRc implies aRc super-sub-type yes or no boundedness whether the hierarchy is bounded – upwards or downwards type-instance downwards bounded or unbounded fixed finite levels fixed or not-fixed number of fixed levels [a number] stratification whether links between non-neighbouring levels are allowed type-instance stratified or unstratified formal generation whether new objects are automatically generated by formal algorithms whole-part fusion yes or no complement yes or no type-instance fusion yes or no super-sub-type fusion yes or no complement yes or no relation class-ness whether the relation is recognised as a first-class object type-instance first- or second-class super-sub-type first- or second-class 46 vertical aspect – visual summary 47 example – stratification Ranks: vertically stratified and unstratified ‘high’ ontologically committing TLOs 48 style.visibility style.visibility example – formal generation two kinds of formal generation {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} relations formally generative fusion complement whole-part yes or no yes or no type-instance yes or no typically, no super-sub-type yes or no yes or no ‘high’ ontologically committing TLOs across three kinds of relation 49 style.visibility style.visibility style.visibility style.visibility framework assessment matrix: vertical aspect 50 horizontal aspect – stratification TLO survey: assessment framework: overall formal structure - framework level 2 0.2 51 horizontal stratification choices {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} Label Unified type Separate types Stratifying relation spacetime spatio-temporal locations spatial locations, temporal locations spaces are multiply located at times (though this is often a derived relation – from an occupying object’s links to both space and time) locations supersubstantival objects (physical) objects, locations objects are (exactly) located at their locations properties objects substances, properties substances are bearers of properties endurants perdurants continuants, occurrents occurrent is dependent upon continuant immaterial (physical) objects material objects, immaterial objects immaterial objects are part of material objects 52 example aspect – spacetime stratification {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} Label Unified type Separate types Stratifying relation spacetime spatio-temporal locations spatial locations, temporal locations spaces are multiply located at times (though this is often a derived relation – from an occupying object’s links to both space and time) revisit this relation in a later diagram 53 style.visibility framework assessment matrix: horizontal aspect The TLOs seem to adopt a general policy of either separating or unifying 54 style.visibility framework assessment matrix: visualizing the horizontal stratification choices can visualize the stratification choices as a journey 55 style.visibility 1 2 3 4 5 6 visualising example 1: BFO stratification journey maximising separation - six strata example of a variant separation: original unified type retained spacetime stratifying relations (time indexed relations suffixed with ‘ … at a time ’) 56 dependence style.visibility style.visibility style.visibility style.visibility visualising example 2: IDEAS stratification journey maximising unification - one strata 57 1 style.visibility universal: individual core commitments TLO survey: assessment framework: framework level 3 58 universal – choices {5C22544A-7EE6-4342-B048-85BDC9FD1C3A} Commitment Choice mereology standard (e.g. GEM) or not interpenetration allowed or not allowed materialism adopted or not adopted possibilia possible worlds or actual world criteria of identity intensional or extensional time presentist or eternalist indexicals: here and now supported - not supported higher arity supported - not supported 59 framework assessment matrix: universal reasonably common choice choice variety advanced choice 60 style.visibility style.visibility style.visibility style.visibility Ontological assessment framework: industry data models 61 ontological assessment framework 62 Assessment framework can be applied to most data models and ontologies Summary TLO Survey: 63 key deliverables we have: identified around forty top-level ontologies developed a framework to assess their ontological commitments assessed the identified top-level ontologies using the framework this provides a solid foundation for an informed selection of a suitable top-level ontology 64 general findings the top levels are very diverse easily seen in the graphical representations common engineering pattern/decision: universal top-level range of levels of ontological commitment there were no existing frameworks for assessing the ontological commitment though some initial work has been done we had to build our own framework, with three levels general formal structure universal 65 the FDM framework looks at a top-level ontology’s ontological commitments in three ways general formal structure universal commitments characterises the choices for the key ontological commitments can be used to: identify the level of ontological commitment made by a top-level ontology profile the top-level ontology’s choices of ontological commitments provides a solid foundation for comparing and selecting a top-level ontology 66 Questions 67