Tutorial Rights and Licenses for Linked Data Part IV ... · Automated RDF Licenses Generation. ESWC...

Tutorial Rights and Licenses for Linked DataPart IV- Applications

Serena Villata

INRIA Sophia Antipolis, France

Licenses in the Web of Data

“the absence of clarity for data consumers about the terms under which theycan reuse a particular dataset, and the absence of common guidelines for

data licensing, are likely to hinder use and reuse of data”Heath and Bizer,

Linked Data: Evolving the Web intoa Global Data Space, 2011


• Support for generating RDF licenses• Share-Alike statements• Licenses compatibility and composition• Open challenges


• Support for generating RDF licensesElena Cabrio, Alessio Palmero Aprosio, Serena Villata: These AreYour Rights - A Natural Language Processing Approach toAutomated RDF Licenses Generation. ESWC 2014: 255-269

• Share-Alike statementsMarkus Krotzsch, Sebastian Speiser: ShareAlike Your Data:Self-referential Usage Policies for the Semantic Web. ISWC 2011:354-369

• Licenses compatibility and compositionGuido Governatori, Antonino Rotolo, Serena Villata, FabienGandon: One License to Compose Them All - A Deontic LogicApproach to Data Licensing on the Web of Data. ISWC 2013:151-166

• Open challenges





• Open challenges





• Open challenges

Support for generating RDF licenses

RESEARCH QUESTIONHow to support users in defining RDF licenses from natural language ones?

@prefix odrl: http://www.w3.org/ns/odrl/2/.@prefix l4lod: http://ns.inria.fr/l4lod/.@prefix : http://example/licenses/.

:licOGL a odrl:Set; odrl:permission [

a odrl:Permission; odrl:action odrl:distribute; odrl:action odrl:derive; odrl:action odrl:commercialize

] ; odrl:duty [ a odrl:Duty; odrl:action odrl:attribute; odrl:action odrl:attachPolicy ] .

Main features

1 RDF representation of licenses - CCRel and ODRLvocabularies,

2 Classification problem in supervised learning - SupportVector Machines,

3 Online service: NLL2RDF(Natural Language License to RDF)

Main features




Main features




Synopsis of the overall framework

NATURAL LANGUAGELICENSES TEXTS

NLL2RDF

RDF LICENSESSPECIFICATION

TOKENIZATION

LEMMATIZATION

PoS TAGGING

PREPROCESSINGMODULE

CLASSIFICATION MODULE

SVM

RDF LICENSES GENERATIONMODULE

NLL2RDF - online demo

Test it!http://www.airpedia.org/nll2rdf-tool/

Share-Alike statements

Goal:• model licenses as part of the data to enable easy

exchange and automated processingSolution:• new policy modeling language to manage Share-Alike

statements

Model of provenance information

Process used Artefact

wasGeneratedBy

Policy

hasPolicy

Derivation

Usage

wasTriggeredBy

PurposehasPurposesubclass of

propertyarrow start: domainarrow end: range

Fig. 1. Informal view of a simple provenance model

sense of “states of the world” such as when an artefact has been published with suit-able attribution), which corresponds to the notion of goal-based policies as defined byKephart and Walsh [19].

To specify the conditions of a policy, we need a model for further describing suchusage processes and their relationships to concrete artefacts. This model in particularmust represent the origin of the artefact, and the context in which it has been published.Such provenance information can be described in various ways, e.g. with a provenancegraph that specifies the dependencies between processes and the artefacts they use andgenerate. Here we use the very simple provenance model illustrated informally in Fig. 1.This base model can of course be further specialised for specific applications and otheruse cases; we just require a minimal setup for our examples.

The provenance model re-uses the vocabulary elements artefact, process, used, was-GeneratedBy, and wasTriggeredBy from the Open Provenance Model. For our partic-ular application, we further split processes into derivations (processes that generate anew artefact) and other usages that only use artefacts without change. To cover the CCuse case, we introduce the hasPurpose property relating a usage to its purpose, e.g.,stating that a usage was non-commercial. The hasPolicy property assigns to an arte-fact a policy, which means that all processes using the artefact are (legally) required tocomply to its policy.

According to OPM, a process p1 wasTriggeredBy another process p2, if p1 can onlyhave started after p2 started. So, somewhat contrary to intuition, the “triggering” israther a precondition but not a necessary cause of the triggered one. A usage restrictionthat requires attribution would thus be formalised as a policy requiring that the usageprocess wasTriggeredBy an attribution process, and not the other way around.

The provenance model provides a basic vocabulary for specifying information aboutartefacts and policies. To realise content-based restrictions we further want to talk aboutthe relationships of policies. For example, ShareAlike requires the value of hasPolicyto refer to a policy which allows exactly the same uses as the given CC SA license.This subsumption between policies is called policy containment, and we introduce apredicate containedIn to express it. Informally speaking, the fact containedIn(p, q) canalso be read as: any process that complies with policy p also complies with policy q.When allowing policy conditions to use containedIn, the question whether or not aprocess complies to a policy in turn depends on the evaluation of containedIn. Ourgoal therefore is to propose a formal semantics that resolves this recursive dependencyin a way that corresponds to our intuitive understanding of the policies that occur inpractice.

Modeling licenses in OWL DL

Public Domain License:• PD : Usage t Derivation.

CC Attribution:• BY : (Usageu∃wasTriggeredBy .Attribution)t (Derivationu∀wasGeneratedBy−1.∀hasPolicy .∃containedIn.{BY}).

Modeling licenses in OWL DL

CC Attribution-NoDerivs:• BY − ND : CBY u (Process u ¬Derivation).

CC Share-Alike:• BY − SA : CBY u∀wasGeneratedBy−1.∀hasPolicy .(∃containedIn.{BY −SA} u ∃containedIn−1.{BY − SA}).

Licenses compatibility andcomposition

QUERYRESULT

QUERYRESULT

QUERYRESULT

Open Government License

Open Database License

CC BY-NC-ND License

What is the license associated to the query result?

??????

RESEARCH QUESTIONS1. How to compose in a compliant way the licensing terms to produce a single composite license? 2. How to produce in an automated way the composite license adopting different composition heuristics?

Main features

1 Combination of Semantic Web languages(machine-readable licenses) - defeasible deontic logic,

2 Extension of existing proposals for licenses compatibilityand composition in service license analysis and CClicenses,

3 Heuristics for licenses combination.

Main features




Main features




Synopsis of the overall framework

CLIENT QUERY

SELECT... WHERE{...}

LICENSESSELECTION

COMPATIBILITY and COMPLIANCE EVALUATION

LICENSESCOMPOSITION

LICENSES COMPATIBILITY AND COMPOSITION MODULE

CLIENT QUERY

QUERY RESULT

SPARQL QUERY RESULT XML + <link URI-Lc>

The formal languageRepresent, and reason about two components:

1 describe ontology of concepts involved in LOD licenses,

2 capture the deontic component of those licenses.

• Rule-based language,

• Ontology rules:regular defeasible logic rules for deriving plain literals,

• a1, . . . ,an ⇒l1c b support the conclusion of b, given a1, . . . ,an,

• Logic of deontic rules:constructive account of basic deontic modalities (obligation,prohibition, permission),

• a,Ob ⇒l2O p: if a is the case and b is obligatory, then Op holds in

license l2.









license l2.









license l2.









license l2.

Composition heuristics

• OR-composition: if at least one of the licenses involved inthe composition owns a clause, then also lc owns it;

• AND-composition: if all the licenses involved in thecomposition own a clause, then also lc owns it;

Composition heuristics

• OR-composition: if at least one of the licenses involved inthe composition owns a clause, then also lc owns it;

• AND-composition: if all the licenses involved in thecomposition own a clause, then also lc owns it;

Proof theory

• Combining licenses,• Checking their compatibility,• Establishing ontology and deontic conclusions which can

be drawn from the composite license,

i.e., if lc = l1 � · · · � ln obtained from l1, . . . , ln thenconclusions derived in the logic are those that hold in theperspective of lc .

Proof theory:Positive definite provability→ in the paper

Proof theory

• Combining licenses,• Checking their compatibility,• Establishing ontology and deontic conclusions which can

be drawn from the composite license,

i.e., if lc = l1 � · · · � ln obtained from l1, . . . , ln thenconclusions derived in the logic are those that hold in theperspective of lc .

Proof theory:Positive definite provability→ in the paper

Proof theory:Positive defeasible provability

Defeasible provability (+∂M lc p):• M lp is a fact; or• there is an applicable strict or defeasible rule r in Rx for

M lp and, for every rule s in Ry for M l ′∼p, eithers discarded orr is weaker than an applicable strict or defeasible rule t inRx for M l′′p.

OR-composition: Rx = Ry is the union set of all rules of alllicenses in the composition

AND-composition: Rx consists of all rules shared by alllicenses in the composition and Ry is the union setof all rules of all licenses in the composition.































Example: l1 and l2 composition

L = {l1, l2}

ROl1 = {r1 :⇒l1O Attribution, r2 :;

l1O Commercial}

ROl2 = {r3 :⇒l2O ∼Commercial, r4 :⇒l2

O ShareAlike, r5 :;l2O Derivative}

OR heuristics for obligationsAND heuristics for permissions

+∂Olc Attribution, +∂Olc ShareAlike, and +∂Plc Derivative

Example: l1 and l2 composition

L = {l1, l2}

ROl1 = {r1 :⇒l1O Attribution, r2 :;

l1O Commercial}

ROl2 = {r3 :⇒l2O ∼Commercial, r4 :⇒l2

O ShareAlike, r5 :;l2O Derivative}

OR heuristics for obligationsAND heuristics for permissions

+∂Olc Attribution, +∂Olc ShareAlike, and +∂Plc Derivative

Evaluation: SPINDle(logic defeasible reasoner)

http://spin.nicta.org.au/spindle/

Real life example:from the logic to SPINdle

F = {Open}L = {lOGL, lODbL, lBY−NC−ND}

ROlOGL= {r1 :⇒lOGL

O Attribution, r2 : Open ;lOGLO Publishing,

r3 : Open ;lOGLO Distribution, r4 : Open ;

lOGLO Derivative,

r5 : Open ;lOGLO Commercial}

ROlODbL= {r6 :⇒lODbL

O ShareAlike, r7 :⇒lODbLO Attribution,

r8 :;lODbLO Sharing, r9 :;

lODbLO Derivative}

ROlBY−NC−ND

= {r10 :⇒lBY−NC−NDO Attribution, r11 :⇒lBY−NC−ND

O ∼Commercial,

r12 :⇒lBY−NC−NDO ∼Derivative, r13 :;

lBY−NC−NDO Sharing}

� = {lODbL � lBY−NC−ND}

Real life example:from the logic to SPINdle

>> Open

r1: =>[Oc]Attribution

r2: Open =>[-Oc] -Publishing

r3: Open =>[-Oc] -Distribution

r4: Open =>[-Oc] -Derivative

r5: Open =>[-Oc] -CommercialExpl

r6: =>[Oc] ShareAlike

r7: =>[Oc] Attribution

r8: =>[-Oc] -Share

r9: =>[-Oc] -Derivative

r10: =>[Oc] Attribution

r11: =>[Oc] -CommercialExpl

r12: =>[Oc] -Derivative

r13: =>[-Oc] -Share

r9 > r12

Real life example:from SPINdle to RDF

AND-composition +∂Olc AttributionOR-composition is admissible: conflict between r5 and r11, and between rule r12 andrules r4 and r9

Deontic conclusions: +∂Olc Attribution, +∂Olc ShareAlike, +∂Plc Publishing,+∂Plc Distribution, +∂Plc Sharing, −∂Plc Derivative, −∂Plc Commercial

SPINdle it takes 14 milliseconds to produce the following conclusions+d [Oc]Attribution,+d [-Oc]-Distribution,+d [-Oc]-Publishing,+d [-Oc]-Share,+d [Oc]ShareAlike















@prefix l4lod: http://ns.inria.fr/l4lod/.

@prefix : http://example/licenses.

:licC a l4lod:License;

l4lod:obliges l4lod:Attribution;

l4lod:obliges l4lod:ShareAlike;

l4lod:permits l4lod:Publishing;

l4lod:permits l4lod:Distribution;

l4lod:permits l4lod:Sharing.



@prefix l4lod: http://ns.inria.fr/l4lod/.

@prefix : http://example/licenses.

:licC a l4lod:License;

l4lod:obliges l4lod:Attribution;

l4lod:obliges l4lod:ShareAlike;

l4lod:permits l4lod:Publishing;

l4lod:permits l4lod:Distribution;

l4lod:permits l4lod:Sharing.

1 Enlarge set of composition heuristics:quantitative ones and Constraining-value

2 Data obtained by inference from one or several licenseddatasets, i.e., queries going beyond basic SELECT queries,where aggregations are present, e.g., average, sum

3 Temporal terms of the licenses4 Licensing vocabularies: meaning, implications, statistics.










Important references

• Gangadharan, Weiss, D’Andrea, Iannella: Service license compositionand compatibility analysis. ICSOC 2007, LNCS 4749. pp. 257-269.

• Gordon: Analyzing open source license compatibility issues withCarneades. ICAIL 2011. pp. 51-55. ACM.

• Miller, Styles, Heath: Open data commons, a license for open data.LDOW 2008.

• Nadah, de Rosnay, Bachimont: Licensing digital content with a genericontology: escaping from the jungle of rights expression languages.ICAIL 2007. pp. 65-69. ACM.

• Pucella, Weissman: A logic for reasoning about digital rights. CSFW2002. pp. 282-294. IEEE.

• Truong, Gangadharan, Comerio, Dustdar, Paoli: On analyzing anddeveloping data contracts in cloud-based data marketplaces. APSCC2011, pp. 174-181. IEEE.

Thanks for your attention!

Date post:	02-Oct-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Tutorial Rights and Licenses for Linked Data Part IV ... · Automated RDF Licenses Generation. ESWC...

Documents