=Paper=
{{Paper
|id=Vol-2377/paper6
|storemode=property
|title=Knowledge Reconciliation with Graph Convolutional Networks: Preliminary Results
|pdfUrl=https://ceur-ws.org/Vol-2377/paper_6.pdf
|volume=Vol-2377
|authors=Pierre Monnin,Chedy Raissi,Amedeo Napoli,Adrien Coulet
|dblpUrl=https://dblp.org/rec/conf/esws/MonninRNC19
}}
==Knowledge Reconciliation with Graph Convolutional Networks: Preliminary Results==
https://ceur-ws.org/Vol-2377/paper_6.pdf
Knowledge Reconciliation with Graph
Convolutional Networks: Preliminary Results?
Pierre Monnin1 , Chedy Raı̈ssi1,2 , Amedeo Napoli1 , and Adrien Coulet1,3
1
Université de Lorraine, CNRS, Inria, LORIA, F-54000 Nancy, France
{pierre.monnin, chedy.raissi, amedeo.napoli, adrien.coulet}@loria.fr
2
Ubisoft, Singapore
3
Stanford Center for Biomedical Informatics Research, Stanford University, 94305
Stanford, California, USA
Abstract. In this article, we investigate the task of identifying nodes
that are identical, more general, or similar within and across knowledge
graphs. This task can be seen as an extension of instance matching or
entity resolution and is here named knowledge reconciliation. In particu-
lar, we explore how Graph Convolutional Networks (GCNs), previously
defined in the literature, can be used for this task and evaluate their per-
formance on a real world use case in the domain of pharmacogenomics
(PGx), which studies how gene variations impact drug responses. PGx
knowledge is represented in the form of n-ary relationships between one
or more genomic variations, drugs, and phenotypes. In a knowledge graph
named PGxLOD, such relationships are available, coming from three dis-
tinct provenances (a reference database, the biomedical literature and
Electronic Health Records). We present and discuss our preliminary at-
tempt to generate graph embeddings with GCNs and to use a simple
distance between embeddings to assess the similarity between relation-
ships. By experimenting on the 68,686 PGx relationships of PGxLOD,
we found that this approach raises several research questions. For ex-
ample, we discuss the use of the semantics associated with knowledge
graphs within GCNs, which is of interest in the considered use case.
Keywords: Knowledge Reconciliation · N -ary relationships · Graph
Embeddings · Graph Convolutional Networks.
1 Introduction
Data and knowledge can be accessed extensively on the Web and interpreted by
both human and software agents. Because these elements of knowledge are of
various provenances, spread in various places and published following distinct
standards, it is challenging to compare and conjointly use their content. Semantic
Web and Linked Open Data (LOD) [2] provide standards and technologies to
?
Supported by the PractiKPharma project, founded by the French National Research
Agency (ANR) under Grant ANR15-CE23-0028, by the IDEX “Lorraine Université
dExcellence” (15-IDEX-0004) and by the Snowball Inria Associate Team.
48 P. Monnin et al.
facilitate the interoperability of knowledge spread over the Web, such as Uniform
Resource Identifiers (URIs) and the Resource Description Framework (RDF)
format. URIs identify nodes that can represent entities of the real world (e.g.,
places, persons, drugs), while RDF statements represent edges, using predicates
to link entities to each others or to literals (e.g., strings, integers). Such predicates
express the semantics of the relationship that connects two nodes or a node and
a literal (e.g., is-born-in, has-firstname). Therefore, URIs and RDF statements
enable to represent knowledge in the form of a directed and labeled multigraph,
loosely called a knowledge graph.
Because datasets are independently published on the Web, possibly with
some overlap, it happens that different URIs are used to identify the same re-
source. For example, dbpedia:Warfarin and wikidata:Q407431 are two URIs
representing the chemical compound Warfarin in DBpedia and Wikidata. As a
consequence, identifying different URIs possibly referring to the same resource is
necessary to use various and initially independent datasets together. This task,
called instance matching [6] or entity resolution, can be extended to identify not
only identical resources but also more general or somehow similar ones, a task we
call knowledge reconciliation (by analogy with reconciliation in databases [1]).
In this work, we illustrate this task with a real world application in the field
of pharmacogenomics (abbreviated PGx). This field studies the influence of ge-
nomic variations in drug response phenotypes. Knowledge in PGx is typically
composed of n-ary relationships between one or more genomic variations, drugs
and phenotypes, stating that a patient having the specified genomic variations,
and being treated with the specified drugs will be more likely to experience the
given phenotypes. PGx relationships can be found in different sources: refer-
ence databases, biomedical literature, or by mining Electronic Health Records
(EHRs). Therefore, there is a need to reconcile these PGx relationships from
different sources, for example to confirm state-of-the-art knowledge found in the
literature with clinical counterpart found in EHRs [5].
Several existing works use Semantic Web technologies to represent PGx
knowledge, as they allow to easily relate resources to other nodes in the knowl-
edge graph that can enrich their semantics (e.g., partOf resources, classes of
ontologies). For example, we built PGxLOD [10], a large knowledge graph con-
taining 68,686 PGx relationships from the three aforementioned sources. As Se-
mantic Web technologies only allow binary predicates, to be represented, PGx
relationships are reified: the relationship itself is a node, linked by predicates
to its components. For example, Figure 1 depicts the reification as the node
pgx rel 1 of a ternary relationship between gene CYP2C9, drug warfarin and
phenotype cardiovascular diseases. A PGx relationship is fully defined by its
components and, accordingly, two relationships involving the same sets of com-
ponents are identical. Hence, reconciliation techniques based on the relational
structure of nodes [6] are well-suited to reconcile PGx relationships represented
using Semantic Web technologies.
In this paper, we investigate how the task of knowledge reconciliation can be
achieved using graph embeddings [4], i.e., low-dimensional vectors representing
Knowledge Reconciliation with GCNs: Preliminary Results 49
CYP2C9 cause
s
causes
pgx rel 1 cardiovascular diseases
s
warfarin cause
Fig. 1. Representation of a reified ternary PGx relationship between gene CYP2C9, drug
warfarin and phenotype cardiovascular diseases. The relationship is reified as the
node pgx rel 1, which is connected to its components by the causes predicate.
graph structures (e.g., nodes, edges, subgraphs) while preserving variously the
properties of the graph. Particularly, we present the preliminary results of an
original experiment using Graph Convolutional Networks (GCNs) [8, 15] that
have already been successfully used for link prediction, a task somehow similar
to knowledge reconciliation. GCNs compute an embedding for each node consid-
ering its neighbors, and, thus, are well adapted to our task in which the relational
structure is of prime importance. Similarity between n-ary relationships could
be represented by ensuring a low distance between their respective embeddings.
Inspired by recent works [13, 16], we use definitions of inverses of predicates
to illustrate how semantics of knowledge graphs could be used in GCNs. We
experimented by reconciling the 68,686 PGx relationships from PGxLOD [10].
The remainder of this article is organized as follows. Section 2 presents related
works. Section 3 details GCNs and the proposed general setting for knowledge
reconciliation of n-ary relationships. Section 4 describes our experiment with the
biomedical knowledge graph PGxLOD. Finally, we discuss our results and future
directions in Section 5.
2 Related Works
Numerous works exist on ontology matching. The interested reader could refer
to [6] for a detailed presentation of approaches. In the following, we focus on
graph embeddings techniques, that have been investigated in multiple works and
successfully applied on knowledge graphs for tasks such as node classification or
link prediction [8, 14, 15]. Works differ in the considered type of graphs (e.g.,
homogeneous graphs, heterogeneous graphs such as knowledge graphs) or in the
graph embedding techniques used (e.g., matrix factorization, deep learning with
or without random walk), as listed in the taxonomies of problems and techniques
in Cai et al. survey [4]. In the following, few specific examples are detailed but
a more thorough overview can be found in some of the existing surveys [4, 11].
A first example is TransE [3], which computes for each triple hs, p, oi of a
knowledge graph, embeddings hs , hp , ho , such that hs + hp ≈ ho , i.e., the
translation vector from the subject to the object of a triple corresponds to the
embedding of the predicate. This approach is adapted for link prediction but,
according to the authors, it is unclear if it can model adequately relationships
50 P. Monnin et al.
of distinct arities, such as 1-to-Many, or Many-to-Many. Another example is
RDF2Vec [14], which first extracts, for each node, a set of sequences of graph
sub-structures starting from this node. Elements in these sequences can be edges,
nodes or even subtrees. Then, sequences feed algorithms from the word2vec neu-
ral language model that compute embeddings for each element in a sequence by
either maximizing the probability of an element given the other elements of the
sequence (Continuous Bag of World architecture) or maximizing the probabil-
ity of the other elements given the considered element (Skip-gram architecture).
A third approach, adopted in this article, is GCNs that have been introduced
in [8] for semi-supervised classification on graphs and extended in [15] for en-
tity classification and link prediction in knowledge graphs. Contrasting TransE
and RDF2Vec that respectively work at the triple and sequence levels, GCNs
compute the embedding of a node by considering its neighborhood. Therefore,
GCNs seem more suited to the task of reconciling n-ary relationships, that are
entirely defined by their neighboring nodes representing their components.
However, previous methods do not consider the semantics associated with
predicates and nodes. Alternatively, Logic Tensor Networks [16] are used to learn
groundings. The grounding of a logical term is a vector of real numbers (i.e., an
embedding) and the grounding of a logical clause is a real number in the interval
[0, 1] (i.e., the confidence in the truth of the clause). The learning process tries
to minimize the satisfiability error of a set of clauses, while trying to ensure the
logical reasoning. This work can interestingly be compared to graph embeddings
if knowledge graphs are considered in their logical form, i.e., considering nodes
as logical terms and edges linking two nodes as logical formulae. We adopted
such consideration by exploring in this work a first manner to include (limited)
semantics within GCNs, for the knowledge reconciliation task.
3 Knowledge Reconciliation with GCNs
3.1 Learning Task
We consider that we have at our disposal a knowledge graph, with specific nodes
representing reified n-ary relationships. Our task consists in learning embeddings
for these relationships such as their distance reflects their similarity or dissim-
ilarity. The learning task relies on two elements: the learning of embeddings
associated with each node of the graph and the assessment of the similarity be-
tween nodes representing n-ary relationships by computing the distance between
their respective embeddings.
To train our GCN model, we constitute a training set and a test set made
of a balanced number of positive and negative examples. Regarding positive
examples, we assume that some n-ary relationships are already labeled as similar
from a manual labeling or from the execution of another method, such as the
similarity rules validated by an expert we use in Section 4. This similarity labels
may have different levels (e.g., very high, high) or may reflect different semantics
(e.g., identical relationships, more general ones). However in this preliminary
setting, we do not take into account such detailed semantics and only consider a
Knowledge Reconciliation with GCNs: Preliminary Results 51
coarse-grained similarity: similar or not labeled. Indeed, as knowledge graphs are
built under the Open World Assumption, absent statements from a knowledge
graph are only unknown and not false. For this reason, we consider that we
have at our disposal only positive similarity labeling for n-ary relationships. To
generate “negative” examples, an approach similar to the one used in TransE [3]
is considered: for each pair of similar n-ary relationships (i, j), another n-ary
relationship k is found such as it is not labeled as similar either to i or to j. The
triple (i, j, k) representing a training example is then added to the training set
S. The same method is used to generate the test set.
3.2 Using GCNs to generate graph embeddings
In the following, we adopt the notations defined in [15]. As such, R denotes
the set of predicates in the considered knowledge graph. Considering a node i
and a predicate r ∈ R, we denote Nir the set of nodes reachable from i by r.
Only nodes and predicates linking nodes are considered. Literals and predicates
linking nodes to literals are discarded.
GCNs can be seen as a message-passing framework of multiple layers, in which
(l+1)
the embedding hi of a node i at layer (l + 1) depends on the embeddings of
its neighbors at level (l), as stated in Equation (1).
!
(l+1)
X X 1
(l) (l) (l) (l)
hi =σ W h + W0 hi (1)
r
ci,r r j
r∈R j∈Ni
The convolution over the neighboring nodes of i is computed with a specific
(l)
weight matrix Wr for each predicate r ∈ R and each layer (l). This convolution
is regularized by a constant ci,r , that can be set for each node and each predicate.
Additionally, to ensure that the embedding of i at layer (l+1) also depends on its
(l)
embedding at layer (l), a self connection is allowed, with the weight matrix W0 .
σ is a non-linear function such as ReLU, used in our experiment (Section 4).
Authors in [15] consider that every predicate r ∈ R has an inverse rinv ∈
R. As this is not always true in knowledge graphs, and to illustrate how the
semantics of predicates could be used in GCNs, we leverage potentially defined
inverse predicates. We consider the three following cases for a predicate r:
(i) If r is defined as symmetric, we do not consider an inverse but ensure that
the adjacency matrix for r is symmetric;
(ii) If r has a defined inverse r−1 , we use r and r−1 and ensure their adjacency
matrices are indeed representing inverse relations;
(iii) Otherwise, we generate an inverse rinv such as its adjacency matrix represent
the inverse of r.
By doing so, we avoid generating an abstract inverse rinv for predicates r having a
defined inverse r−1 or being symmetric, which would add unnecessary messages.
Indeed, consider a predicate r, its defined inverse r−1 , and two nodes i and
r r −1
j such that edges i −
→ j and j −−→ i are in the knowledge graph. By always
52 P. Monnin et al.
r r −1
generating an abstract inverse, two edges would be added, j −−inv −→ i and i −−inv
−→ j,
duplicating the existing edges and adding two unnecessary messages.
To train GCNs, we minimize the loss function presented in Equation (2),
inspired from the one used in TransE [3].
X
L= max ||hi − hj ||2 + γ − ||hi − hk ||2 , 0 (2)
(i,j,k)∈S
Given a training example (i, j, k) from the training set S, minimizing the loss
function aims at minimizing the distance between hi and hj , i.e., ensuring their
similarity, while maximizing the distance between hi and hk . The constant γ is
a margin hyperparameter aiming at increasing the difference between the two
distances.
4 Experimentation on PGx Knowledge
4.1 Input Knowledge Graph: PGxLOD
We experimented our approach on PGxLOD [10], a large knowledge graph
containing PGx relationships from three distinct sources: PharmGKB (a refer-
ence database), the biomedical literature and results from studies on Electronic
Health Records. Main statistics of PGxLOD are presented in Table 1.
Table 1. Main statistics of PGxLOD. The line “Predicates” only counts predicates
used to link two nodes together, excluding literals. As our experiment uses a 3-layer
network, only the 3-hop neighborhood of PGx relationships is considered to compute
their embeddings.
Triples 59,136,400
Nodes and literals 25,226,599
Predicates 378
PGx relationships 68,686
Nodes in their 3-hop neighborhood 2,943,613
Edges in their 3-hop neighborhood 32,773,429
Similarity links between PGx relationships 283,248
owl:sameAs links 109,226
skos:broadMatch links 136,264
skos:relatedMatch links 37,758
In PGxLOD, some similar PGx relationships are already labeled by being
linked together with one of the three following predicates: owl:sameAs express-
ing identical relationships, skos:broadMatch expressing more general ones and
skos:relatedMatch expressing related ones to some extent. Such labels result
from the application of logical reconciliation rules that are described in [10].
Knowledge Reconciliation with GCNs: Preliminary Results 53
They are used to constitute the training and test sets, in a “knowledge graph
as silver standard” perspective [12]. Even if owl:sameAs, skos:broadMatch and
skos:relatedMatch links express different similarity semantics, here we indif-
ferently consider the three predicates as expressing a coarse-grained similarity.
Thus, two PGx relationships are labeled as similar if they are linked by one
of these three predicates. Additionally, we consider their adjacencies in an un-
directed perspective, i.e., having (i, j) as a similarity edge is equivalent to having
(j, i). As owl:sameAs and skos:relatedMatch are symmetric, numbers of avail-
able links for training and test sets are consequently half of those presented in
Table 1. For each predicate, the training set is constituted by 32 of the links and
the test set by 13 . To form triples (i, j, k) in these sets, k is chosen such as it is
not directly linked via a similarity predicate either to i or to j.
4.2 Experimental Setting
For our preliminary experiment, we use a standard architecture and hyperpa-
rameters previously reported in the literature with successful uses of GCNs. We
consider a 3-layer network where each layer uses ReLU as the activation func-
tion. In such a 3-layer architecture, only neighboring nodes up to 3 hops of PGx
relationships will have an impact on their embeddings, output at layer 3 (Equa-
tion (1)). The input layer consists in a featureless approach as in [8, 15], i.e., the
input is just a one-hot vector for each node of the graph. Both the input layer
and the hidden layer have an output dimension of 16 while the output layer has
an output dimension of 10. Therefore, embeddings for all nodes in the knowl-
edge graph are in R10 . Only the embeddings for nodes representing the reified
PGx relationships are of interest in our reconciliation task and are considered
in the loss function. As in [15], the constant ci,r is set to |Nir | and we use the
basis-decomposition with 10 basis to avoid the growth in number of parameters.
For the learning process, we use the Adam optimizer [7] with a starting learning
rate of 0.01 and a L2-regularization coefficient of 0.0005. The margin hyperpa-
rameter γ is set to 2. Our experiment was implemented using PyTorch and the
Deep Graph Library.
4.3 Results
We trained our model during 60 epochs. The last layer outputs embeddings for
all nodes in the graph but we only consider the ones representing reified PGx re-
lationships. The mean and standard deviation of distances between embeddings
of similar relationships in the training set were respectively µtrain = 1.93 and
σtrain = 4.18. As a simple evaluation, for each example (i, j, k) from the test set,
(i, j) or (i, k) were considered as similar if their embeddings were distant of less
than µtrain + σtrain . These results were compared with existing similarity links,
obtaining a precision of 0.92, a recall of 0.94 and a F1-score of 0.93.
Then, we investigated differences between the three similarity predicates. 2D
projections of embeddings using UMAP [9] are depicted in Figure 2. We can see
54 P. Monnin et al.
that clusters of nodes are appearing but seem still close. More epochs or a wider
neighborhood may allow to emphasize the difference between such clusters.
15 10
10
10 5
5
5
0
0
0
5
5
5
10 10
10
10 5 0 5 10 15 10 5 0 5 10 15 15 10 5 0 5
(a) sameAs links (b) broadMatch links (c) relatedMatch links
Fig. 2. 2D projections using UMAP [9] of embeddings of PGx relationships involved
in the given similarity predicates.
Figure 3 depicts the distributions of distances between embeddings of similar
PGx relationships depending on their similarity link. The low means for the three
predicates illustrate that similar relationships have indeed embeddings with low
distances. We notice that the mean distances for owl:sameAs are the lowest
while the ones for skos:relatedMatch are the greatest. This could indicate the
ability of the network to learn the close similarity expressed by owl:sameAs
links and the more fuzzy one expressed by skos:relatedMatch links. Regarding
skos:broadMatch links, the distance ranges and variances are more important
than for the two other predicates. This could also illustrate the ability to learn the
semantics of the predicate. Additionally, skos:broadMatch links are directed,
and thus, could be more difficult to fit correctly, mixed with symmetric similarity
predicates. Also, only skos:broadMatch links connect PGx relationships across
the three considered sources [10]. Therefore, they link together relationships that
may have more diversity in the semantics and vocabularies of their components,
potentially explaining the higher distance ranges and variances.
sameas training set 105 broadmatch training set 104 relatedmatch training set
104 sameas test set broadmatch test set relatedmatch test set
104
103
103
103
102 102
102
101 101
101
100 100 100
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 50 100 150 200 250 300 350 0 20 40 60 80
(a) sameAs links (b) broadMatch links (c) relatedMatch links
µtrain = 0.33 µtrain = 2.27 µtrain = 4.16
σtrain = 0.26 σtrain = 4.90 σtrain = 2.77
µtest = 0.33 µtest = 2.30 µtest = 4.35
σtest = 0.26 σtest = 5.12 σtest = 3.31
Fig. 3. Distributions of distances between embeddings of similar PGx relationships
linked by the given similarity predicates. µ and σ respectively denote the mean and
the standard deviation.
Knowledge Reconciliation with GCNs: Preliminary Results 55
5 Discussion and Conclusion
In this paper, we investigated the task of generating graph embeddings for knowl-
edge reconciliation using Graph Convolutional Networks and ensuring that em-
beddings associated with similar reified n-ary relationships have a low distance.
We experimented our approach on the real world use case of reconciling PGx
n-ary relationships from three distinct sources. Our preliminary results found
this approach to be suitable and to raise several research questions.
First, the network output different distances for the three considered predi-
cates: owl:sameAs, skos:broadMatch and skos:relatedMatch. This could indi-
cate that it was able to learn their different similarity semantics or had difficulties
to adequately fit some of them. Possible improvements would be to increase the
number of epochs or test other values for hyperparameters. The loss function
could integrate the different semantics of the predicates linking similar rela-
tionships i and j, for example by considering three different weighted sums.
Separate models could be learned: one per predicate or one for owl:sameAs and
skos:relatedMatch and another for skos:broadMatch which is not symmetric,
which could make easier interpreting the semantics of the output similarity.
Because we used a 3-layer network, only nodes in the 3-hop neighborhood
of PGx relationships were considered for the computation of their embeddings.
However, nodes in further neighborhoods may bring additional semantics. In
particular, phenotypes extracted from the biomedical literature are frequently
complex and formed by several simpler phenotypes, indicated by dependsOn
links. Therefore, we could benefit from using a network with more layers.
We also illustrated how semantics associated with a knowledge graph can be
used in GCNs by considering the definitions of inverses of predicates. This could
be improved, for example by considering the semantics of owl:sameAs links
between nodes. Indeed, these links indicate identical nodes that are currently
considered as neighboring nodes and used as such in the embeddings compu-
tation. Thus, a pre-processing step could consist in mapping nodes linked by
owl:sameAs links into a unique node. Additionally, the generation of negative
examples could be improved by considering ontologies. In such case, PGx rela-
tionships whose components instantiate classes in different parts of an ontology
could be more interesting negative examples.
Our model was evaluated using a manually-defined threshold on distances
between embeddings of relationships to assess their similarity. Other methods
such as (multi-)classification machine learning models could also be investigated.
Advanced performance results could also be computed on knowledge graphs from
other domains as well as be compared with other state-of-the-art methods pre-
sented in Section 2. Finally, we should manually check on a few examples whether
relationships considered as close given the distance between their embeddings
but not linked by any of the similarity predicates are indeed similar.
To conclude, these results constitute solely an initial attempt to use graph
embeddings for the non-trivial task of reconciling n-ary relationships. Several
future directions are considered, among which is the further integration of the
semantics associated with knowledge graphs in GCNs.
56 P. Monnin et al.
References
1. Abiteboul, S., Manolescu, I., Rigaux, P., Rousset, M., Senellart, P.: Web Data
Management. Cambridge University Press (2011)
2. Bizer, C., Heath, T., Berners-Lee, T.: Linked data - the story so far. Int. J. Semantic
Web Inf. Syst. 5(3), 1–22 (2009)
3. Bordes, A., Usunier, N., Garcı́a-Durán, A., Weston, J., Yakhnenko, O.: Translating
embeddings for modeling multi-relational data. In: Advances in Neural Information
Processing Systems 26: 27th Annual Conference on Neural Information Processing
Systems 2013. Proceedings of a meeting held December 5-8, 2013, Lake Tahoe,
Nevada, United States. pp. 2787–2795 (2013)
4. Cai, H., Zheng, V.W., Chang, K.C.: A comprehensive survey of graph embedding:
Problems, techniques, and applications. IEEE Trans. Knowl. Data Eng. 30(9),
1616–1637 (2018)
5. Coulet, A., Smaı̈l-Tabbone, M.: Mining electronic health records to validate knowl-
edge in pharmacogenomics. ERCIM News 2016(104) (2016)
6. Euzenat, J., Shvaiko, P.: Ontology Matching, Second Edition. Springer (2013)
7. Kingma, D.P., Ba, J.: Adam: A method for stochastic optimization. CoRR
abs/1412.6980 (2014), http://arxiv.org/abs/1412.6980
8. Kipf, T.N., Welling, M.: Semi-supervised classification with graph convolutional
networks. CoRR abs/1609.02907 (2016), http://arxiv.org/abs/1609.02907
9. McInnes, L., Healy, J., Saul, N., Grossberger, L.: Umap: Uniform manifold approx-
imation and projection. The Journal of Open Source Software 3(29), 861 (2018)
10. Monnin, P., Legrand, J., Husson, G., Ringot, P., Tchechmedjiev, A., Jonquet, C.,
Napoli, A., Coulet, A.: PGxO and PGxLOD: a reconciliation of pharmacogenomic
knowledge of various provenances, enabling further comparison. BMC Bioinfor-
matics 20-S(4), 139:1–139:16 (2019)
11. Nickel, M., Murphy, K., Tresp, V., Gabrilovich, E.: A review of relational machine
learning for knowledge graphs: From multi-relational link prediction to automated
knowledge graph construction. Proceedings of the IEEE 104(1), 11–33 (2016)
12. Paulheim, H.: Knowledge graph refinement: A survey of approaches and evaluation
methods. Semantic Web 8(3), 489–508 (2017)
13. Paulheim, H.: Make embeddings semantic again! In: Proceedings of the ISWC
2018 Posters & Demonstrations, Industry and Blue Sky Ideas Tracks co-located
with 17th International Semantic Web Conference (ISWC 2018), Monterey, USA,
October 8th - to - 12th, 2018. (2018)
14. Ristoski, P., Paulheim, H.: Rdf2vec: RDF graph embeddings for data mining. In:
The Semantic Web - ISWC 2016 - 15th International Semantic Web Conference,
Kobe, Japan, October 17-21, 2016, Proceedings, Part I. pp. 498–514 (2016)
15. Schlichtkrull, M.S., Kipf, T.N., Bloem, P., van den Berg, R., Titov, I., Welling,
M.: Modeling relational data with graph convolutional networks. In: The Semantic
Web - 15th International Conference, ESWC 2018, Heraklion, Crete, Greece, June
3-7, 2018, Proceedings. pp. 593–607 (2018)
16. Serafini, L., d’Avila Garcez, A.S.: Learning and reasoning with logic tensor net-
works. In: AI*IA 2016: Advances in Artificial Intelligence - XVth International
Conference of the Italian Association for Artificial Intelligence, Genova, Italy,
November 29 - December 1, 2016, Proceedings. pp. 334–348 (2016)