This paper presents the results of the ALOD2Vec Matcher in the Ontology Alignment Evaluation Initiative (OAEI) 2020. The matching system exploits a Web-scale dataset, i.e. WebIsALOD, as background knowledge source. In order to make use of the dataset, the RDF2Vec approach is applied to derive embeddings for each concept available in the dataset. ALOD2Vec Matcher participated in the OAEI 2018 campaign before. This is the system's second participation. The matching system has been extended, improved, and achieves better results this year.3 The ALOD2Vec Matcher is an element-level, label-based matcher which uses a large-scale Web-crawled RDF dataset of hypernymy relations as general purpose background knowledge. The dataset contains many tail-entities as well as instance data such as persons or places which cannot be found in common thesauri. In order to exploit the external dataset, a neural language model approach is used to obtain a vector for each concept contained in the dataset. This matching system system was initially introduced at the OAEI 2018 [14] and has been completely re-implemented. The implementation is now based on the Matching EvaLuation Toolkit [5,11] as well as the KGvec2go [12] REST API. A contribution of this paper is also an extension to the MELT framework in the form of a KGvec2go Java client available in the MELT-ML module [6] of MELT 2.6.
1.1
1.2
After the basic concepts of this matcher are introduced (Foundations ), the speci c techniques applied are presented. 3 Copyright c 2020 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
WebIsALOD Dataset A frequent problem that occurs when working with
external background knowledge is the fact that less common entities are not contained
within a knowledge base. The WebIsA [
RDF2Vec The background dataset can be viewed as a very large knowledge
graph; in order to obtain a similarity score for nodes and edges in that graph,
the RDF2Vec [
5 see http://webisa.webdatacommons.org/concept/european_union_ 6 see http://webisa.webdatacommons.org/ 7 For a good overview of the RDF2Vec approach and its applications, refer to http://www.rdf2vec.org/ 8 see http://kgvec2go.org/api.html
The rst step is to link the obtained labels from the ontology to concepts in
the WebIsALOD dataset. Therefore, string operations are performed on the label
and it is checked whether the label is available in WebIsALOD. If it cannot be
found, a token-lookup is performed. Given two entities e1 and e2, the matcher
uses their textual labels to link them to concepts e01 and e02 in the external
dataset. Afterwards, the embedding vectors ve01 and ve02 of the linked concepts
(e01 and e02) are retrieved via a Web request and the cosine similarity between
those is calculated. Hence: sim(e1; e2) = simcosine(ve01 ; ve02 ). If sim(e1; e2) > t
where t is a threshold in the range of 0 and 1, a correspondence is added to a
temporary alignment. In a last step, a one-to-one arity is enforced by applying
a Maximum Weight Bipartite [
In order to consume the vectors in Java, a client has been implemented and contributed to the MELT-ML module. The KGvec2go REST API can now be accessed though class KGvec2goClient. Even though this matcher only uses the WebIsALOD dataset, the implementation supports all datasets accessible on KGvec2go. The extension is available by default in MELT 2.6.
Instance Matching For the 2020 version of the matching system, an instance matching module has been added. After classes and properties have been matched, instances are matched using a string index. The con dence score assigned to instances belonging to matched classes is higher than that of matches between instances belonging to non-matched classes.
Explainability ALOD2Vec Matcher provices an explanation for every
correspondence that is added to the nal alignment. Therefore, the extension
capabilities of the alignment format [
The 2020 system has been completely rewritten. Among the signi cant changes
are an improved handling of string matches, an instance matching module for
the knowledge graph track [
2.1
On the anatomy dataset, the recall could be signi cantly improved in 2020 compared to the 2018 version of the matching system. Despite a drop in precision, the new ALOD2Vec Matcher achieves an overall higher F1 score. Due to multiple API calls to KGvec2go, the runtime performance decreased compared to the 2018 version of the matcher. 2.2
On the conference track, the new matcher con guration achieved a better result than the 2018 one in terms of F1 due to a higher recall (from 0.5 in 2018 to 0.52 in 2020). The overall F1 score on ra1-M3 was 0.59.
10 The 2018 version of the matching system had to be submitted via a download link due to its large size. The 2020 version was submitted using the default process. 11 see https://github.com/janothan/ALOD2VecMatcher
This is the rst year that ALOD2Vec Matcher participates in the knowledge
graph track. The system could complete all matching tasks in time. Due to the
new instance matching module, this matcher obtains the second best results
achieving almost the same score as the Wiktionary Matcher 2020 [
In this paper, we presented the newest version of the ALOD2Vec Matcher, a matcher utilizing an RDF2Vec vector representation of the WebIsALOD dataset, as well as its results in the 2020 OAEI. The matching system has been improved compared to its 2018 version. ALOD2Vec Matcher now uses a remote vector API which makes the matcher package very portable due to its substantially reduced size. Overall, the results of the matching system could be signi cantly improved compared to its last OAEI participation and is the second best performing system on the knowledge graph track.