An important part of human history and knowledge is made of events, which can be aggregated and connected to create stories, be they real or fictional. These events and the stories created from them can typically be inherently complex, reflect societal or political stances and be perceived diferently across the world population. The Semantic Web ofers technologies and methods to represent these events and stories, as well as to interpret the knowledge encoded into graphs and use it for diferent applications, spanning from narrative understanding and generation to fact-checking.
Representing and instantiating events has always been a crucial task for the Semantic Web
community, with some relevant contributions such as specialised ontologies [
The workshop has been held on May 29th, 2023, opened by a keynote talk “Video, Narratives
and Knowledge Graphs” by our invited speaker Paul Groth, Professor of Algorithmic Data
Science at the University of Amsterdam where he leads the Intelligent Data Engineering Lab
(INDElab). He presented his work on entity and event discovery in captioned videos [
In the session KGs for Understanding Events and Stories, Kozaki Kouji et al. introduced the Datasets of Mystery Stories for Knowledge Graph Reasoning Challenge, based on knowledge graphs extracted from Sherlock Holmes’s stories. Franz Krause et al. presented their work for identifying and understanding semantic transitions, titled On the Combination of Event Calculus and Empirical Semantic Drifts . In Musical Meetups: a Knowledge Graph approach for Historical Social Network Analysis, the MEETUPS ontology and knowledge graph have been presented by Alba Morales Tirado et al. The session Event Information Extraction started with the presentation by Heikki Rantala et al., in which connections between places, people and events are extracted in Finding and explaining relations in a biographical knowledge graph based on life events: Case BiographySampo. GPT-3 has been used to realise an event relation dataset in Prompt-based Data Augmentation for Semantically-precise Event Relation Classification by Youssra Rebboud et al. Lars Michaelis’ work titled WikiEvents - A Novel Resource for NLP Downstream Tasks introduces a KG for of event-related location extraction and entity linking. Finally, a Comprehensive Survey on Ontologies about Event has been proposed by Rajesh Piryani et al., making the first step towards the realisation of a 5W1H-compliant ontology.
The workshop attracted over 40 attendees in this first edition. Details about the workshop, including the Program Committee, are available at https://anr-kflow.github.io/semmes/.
The authors are very grateful to Lise Stork, who largely contributed to organise SEMMES, as well as all the Program Committee. This workshop has been partially supported by the French National Research Agency (ANR) within the kFLOW project (Grant n°ANR-21-CE23-0028) and by EU H2020 under the Marie Skłodowska-Curie grant agreement no 812997 (CLEOPATRA).