Machine learning and knowledge graphs are currently essential technologies for designing and building large scale distributed intelligent systems. Machine learning is a well established field, which has currently gained a high momentum due to the advances in the computational infrastructures, availability of Big Data, and appearance of new techniques such as deep learning. In fact, Deep Learning methods have become an important area of research, achieving some important breakthrough in various research fields, especially Natural Language Processing (NLP), Image and Speech Recognition.
Knowledge Graphs are large networks of real-world entities described in terms of their semantic types and their relationships to each other. Knowledge graphs is a recent technology with very high practical impact: examples include the Google Knowledge Graph with over 70 billion facts (in 2016), dataCommons, DBPedia, YAGO and YAGO2, Wikidata and Knowledge Vault, a very large scale probabilistic knowledge graph created with information extraction methods for unstructured or semi-structured information. Specifically, Knowledge Graphs provide the means of development of the newest methods for data management, data fusion / data merging, and graph and network optimization and modeling, serving as a source of high quality data and a base for ubiquitous information integration. While Knowledge Graphs provide explicit knowledge representations in terms of underlying ontologies based on symbolic logic, machine learning, such as Deep Learning technologies, provide implicit or latent representations of the knowledge contained in their models.
Although machine learning (and particularly, recently deep learning) and Knowledge Graphs technologies have been deployed separately, in the last years, the first works combining these technologies are showing large potential in solving many real-world challenges.
In order to pursue more advanced methodologies, it has become critical that the communities related to Machine Learning, Deep Learning and Knowledge Graphs join their forces in order to develop more effective algorithms and applications. In particular, two main technology directions solicited for this special issue are as follows:
i) Improved Machine Learning with Knowledge Graphs: employing semantic models and linked data for the training steps, learning effective representation from the Knowledge Graphs for the tasks of feature extraction, classification, prediction and decision making. A successful example here includes IBM Watson question answering system, that has outperformed the best human players of an intellectual TV quiz show \”jeopardy!\”
ii) Machine Learning for Improving Knowledge Graphs: while capturing semantics correctly is impossible without at least some human involvement, machine learning can assist the knowledge acquisition of the semantic structures substantially. For example, knowledge graphs can be created by employing Deep learning, and then subsequently verified by the humans.
A feature that explains the learnt knowledge graphs to the human needs to be inbuilt in the learning and verification process, making the resulting knowledge graph design solutions explainable to humans. Thus, the knowledge graphs created in this manner may be more efficient and provide more insights than the ones created solely by humans.
The solicited contributions may comprise an overview of the outlined fields, and new technical solutions applicable to such use cases as manufacturing / smart factory, production 4.0 and its control of quality, mobility, smart cities, smart homes and buildings, energy efficiency, health and wellbeing, life sciences, libraries and archives, art.
Such use cases require both the analysis of data (e.g. sensor readings, distributed data processing) – i.e. quantitative information, and as well as qualitative (e.g. information semantically defining the production quality criteria and what is considered to be the deviation from them). As physical infrastructures and architectures (stemming from the fields of Grid and Cloud Computing, Big Data and Internet of Things), have a very high impact on the solutions to be suggested, we consider that Future Generation Computer Systems is the perfect venue for publishing this special issue, given the research area of the journal.
Topics of Interests:
We invite submission of high-quality manuscripts reporting relevant research in the area of generation of knowledge graphs by using deep learning techniques. Topics of interest include, but are not limited to:
Architectures for systems based on Machine Learning and Knowledge Graphs
Machine Learning and Knowledge Graphs in distributed large scale systems
Information management with on Machine Learning and Knowledge Graphs
Probabilistic Knowledge Graphs
Creation and maintenance (curation, quality assurance…) of Knowledge Graphs employing Machine Learning
Machine Learning techniques employing Knowledge Graphs
Explainable Artificial Intelligence for data-intensive systems based on Knowledge Graphs
Non-functional application of Knowledge Graphs in data-intensive systems e.g. for legal use of data, user consent solicitation, smart contracting, GDPR
New approaches for combining Deep Learning and Knowledge Graphs
Methods for generating Knowledge Graph (node) embeddings
Temporal Knowledge Graph Embeddings
Applications of combining Deep Learning and Knowledge Graphs
Recommender Systems leveraging Knowledge Graphs
Link Prediction and completing KGs
Ontology Learning and Matching exploiting Knowledge Graph-Based Embeddings
Knowledge Graph-Based Sentiment Analysis
Natural Language Understanding/Machine Reading
Question Answering exploiting Knowledge Graphs and Deep Learning
Trend Prediction based on Knowledge Graphs Embeddings
Domain Specific Knowledge Graphs (e.g. Smart Cities, Scholarly, Biomedical, Musical)
Applying knowledge graph embeddings to real world scenarios.
Paper submission due: 30 November, 2019
Initial review feedback: 31 January, 2020
Revision Due: 31 March, 2020
Final review decision: 31 July, 2020
Manuscripts must be submitted through the electronic submission system, and please choose \”VSI: ML & KGraphs\”.
Papers will be evaluated for their originality, contribution significance, soundness, clarity, and overall quality. The interest of contributions will be assessed in terms of technical and scientific findings, contribution to the knowledge and understanding of the problem, methodological advancements, and/or applicative value.
– Mehwish Alam, FIZ Karlsruhe – Leibniz Institute for Information Infrastructure, firstname.lastname@example.org
– Anna Fensel, University of Innsbruck, email@example.com
– Jorge Martinez-Gil, Software Competence Center Hagenberg GmbH, firstname.lastname@example.org
– Bernhard A. Moser, Software Competence Center Hagenberg GmbH, email@example.com
– Diego Reforgiato Recupero, University of Cagliari, firstname.lastname@example.org
– Harald Sack, FIZ Karlsruhe – Leibniz Institute for Information Infrastructure, email@example.com