It’s Time We Give Each Other More Data Spaces: Impressions from the Pre-conference Day at SEMANTiCS 2023

This year’s pre-conference day of SEMANTiCS – the annual European conference on semantic technologies, organized by our partners at the Semantic Web Company (SWC), was dedicated to DBpedia. It also featured workshops, covering FAIR Ontologies, Semantic Data Space For Transport, and Natural Language Processing (NLP) for knowledge graph construction.

The day opened with a DBpedia Keynote Session by Edward Curry, Professor of Data Science at the University of Galway. The talk focused on the importance (and the needed research agenda) of introducing Foundational Data Space Models as an enabling platform for data spaces. Edward Curry highlighted the emergence of data management challenges and the need for a unified lifecycle for data spaces and foundational models. Curry listed knowledge graphs, search, and language models as the main pillars of the evolution of data and data spaces.

Ed Curry @EdwardACurry shows the current challenges: we have vast amount of data and our technologies for handling such a deluge are lagging back #semanticsconf @SemanticsConf pic.twitter.com/vNgdIJQtdk

— Angelo A. Salatino (@angelosalatino) September 20, 2023

The event continued with a workshop organized by DatabriX, accompanied by a panel of experts discussing syntactic and semantic interoperability. Individual sessions were dedicated to Environmental – Social – Corporate Governance (ESG) data, Knowledge Engineering of Taxonomies and Ontologies, scholarly communication, and NLP in Healthcare. 

Data Spaces Matter

Although busy preparing Ontotext’s booth and our talk on Large Language Models (LLMS) and knowledge graphs for the opening day of the conference, I managed to attend the panel discussion on Interoperability in Data Spaces. The panelists were Phil Archer (GS1, formerly W3C), Edward Curry, Felix Sasaki (SAP), and Silvia Castellvi (International Data Spaces Association), and were moderated by Vassilios Peristeras. 

Panel Discussion: Interoperability in Data Space, featuring (from left to right) Felix Sasaki (SAP), Phil Archer (GS1, formerly W3C), Edward Curry, Silvia Castellvi (International Data Spaces Association), and Vassilios Peristeras at SEMANTICs 2023

The Current State of Data Spaces and the Future of Interoperability

With forthright questions by Vassilios Peristeras, whose own teaching and research interests gravitate around business information systems, eGovernment, eParticipation, eCommerce, interoperability, and Open and Linked Data, the panel equipped us with several good perspectives. 

Panelists underlined the complexity and significance of data- and many times policy-related challenges and the need for collaborative, innovative approaches to address them. According to them, the main challenges in the emerging common European data space environment were the lack of incentives for data publishers to share their data, the lack of tools to empower and democratize the process of sharing data, and the lack of trust in the system. All agreed that the case of schema.org or the HTML protocol proved that people needed incentives to publish standard data. At the same time, they talked about the fine line between a standard that is an enabler and a standard that is a barrier.

The panel ended with a word from each panelist, among which Edward Curry’s observation struck a chord:

The point of semantics is to continue the journey! It is connected directly with digitalization. And there is no endpoint, it is a continual process, you need to engage with it constantly.

ORKG: Taming Academic Text

One of the presentations I couldn’t afford to not check was the one about the Open Research Knowledge Graph Project.

I have been following  ORKG since its first public announcement. It was back in 2019 when Dr. Sören Auer (Director of TIB) gave a talk titled: Towards Knowledge Graph-based Representation, Augmentation and Exploration of Scholarly Communications. It provided an overview of the ORKG initiative, which aims to describe research papers in a structured manner and make them easier to find and compare.

Anna-Lena Lorenz and Jennifer D’Souza presenting the Open Research Knowledge Graph (ORKG) at SEMANTICs 2023

In the tutorial, Anna-Lena Lorenz and Jennifer D’Souza showed the platform, sharing how it was curated by people and informed by the so-called ORKG Observatories. These are hubs that can organize research contributions (curating papers and knowledge in the respective field)  for anyone interested in having their domain represented in the project. They also demonstrated how the platform is used in practice, how taxonomies are incorporated into the workflow, and how curating content relies on crowd-sourced work together with NLP techniques.

Quo Vadis Linked Data?

Another inspiring part of the pre-conference day, and SEMANTiCS as a whole, was the Academia Meets business. It reinforced the organizer’s long-standing reputation for bridging the gap between academic research and practical industry applications and fostering an environment for sharing knowledge and hatching ideas.

#SemanticsConf Update: Academia and Business came together yesterday in order to discuss and explore challenges & innovative ideas in the #SemanticWeb field. #KnowledgeGraphs, #LLM and #Semantics were the main topics addressed. #Networking @GfWM pic.twitter.com/5Olwskwyhg

— SEMANTiCS Conference (@SemanticsConf) September 21, 2023

As a great epilogue of the pre-conference day and its themes, the session offered a dialogic arena for major topics, chosen from the organizers based on the accepted papers and presentations. And, while I didn’t get there, the artifacts of the dialogues found me in one of the conference halls and one more time proved that “we can only make progress when we join our expertise and passion.“ 

The artifacts of the lively session Academia Meets Business at SEMANTiCS 2023

The day ended with a beer in an impressive setting at Beyerhaus and in anticipation of the next day packed with LLMs and knowledge graph talks.

Stay tuned for my report from Day 1, which was all about the interplay between LLMs and knowledge graphs!