Machine Learning (ML) methods and techniques are playing an increasingly important role for the Semantic Web showing a valuable role for solving tasks such as knowledge extraction, entity (inter-)linking, rule mining, entity retrieval, recommendation and more recently, representation learning and embeddings. In particular, with the wide availability of large amounts of structured and/or semi-structured data on the Web, machine learning techniques have been essential for extracting, fusing and integrating data and knowledge, for instance, for knowledge graph construction or knowledge base augmentation. In addition, neural networks and deep learning-based approaches for distributional semantics and word/entity embeddings have gained wide popularity for a range of tasks, many of which are increasingly coming to rely on representation learning for state-of-the-art performance.
This track invites high quality submissions that focus on the definition, adoption, use and assessment of ML methods for the SW. Papers showing how ML methods and techniques have been applied to develop and improve the SW field and Linked Open Data, or showing how SW technologies and resources have been used to enhance different ML tasks are also welcome. Reproducibility of all presented works is strictly required and should ideally be supported by providing links to used datasets, source code, queries/test cases or live deployments.
Topic of Interest
Topics of interest include but are not limited to:
- Machine learning for knowledge graph construction, completion, refinement
- Statistical relational learning for the Web of data, knowledge graphs, ontologies
- Data mining and knowledge discovery from knowledge bases, linked data and ontologies
- Learning of knowledge graph and entity embeddings
- Machine learning for knowledge and information extraction, for instance, named entity disambiguation, sentiment analysis, relation extraction, or the detection of claims, facts and stances from unstructured documents
- Representation learning for the integration of unstructured data with knowledge bases
- Machine learning for ontology matching, instance matching, search and retrieval
- Learning from big data for large-scale knowledge graphs
- Supervised techniques for matching, fusing or interpreting entity-centric Web markup (e.g. schema.org) or Web tables
- Machine learning-based approaches towards question answering
- Machine learning to augment, fuse or retrieve structured data and knowledge graphs
- Machine Learning approaches like transfer learning, deep learning, graph-kernels, tensor methods, relational graphical models for the Semantic Web.
- Representation Learning techniques for knowledge graphs and linked data, especially hybrid models that combine text, graph structure and semantics
Stefan Dietze, GESIS & Heinrich-Heine-University Düsseldorf, Germany
Mayank Kejriwal, University of Southern California Information Sciences Institute, USA