Insights and Highlights from the ISWC’23 Conference
We are excited to share our experiences and insights from the recent International Semantic Web Conference (ISWC), a premier venue for discussing the latest advancements in semantic web technologies. Our team had the privilege of not only attending but also contributing significantly to the conference.
Workshop Paper Presentation by Majlinda and Fajar
Our team members, Majlinda Llugiqi and Fajar J. Ekaputra, presented a workshop paper at the LM-KBC (Language Model Knowledge Base Construction) Challenge. Their paper, titled “Knowledge-centric Prompt Composition for Knowledge Base Construction from Pre-trained Language Models,” is a paper in advancing the field of knowledge base construction. Majlinda and Fajar’s methodology enhances prompt context by integrating training data and detailed relation descriptions, showcasing the versatility of language models in knowledge base construction.
Guntur’s Paper Presentation
Guntur presented a paper titled “Towards A Knowledge Graph-based Exploratory Search for Privacy Engineering” at the 8th Visualization and Interaction for Ontologies, Linked Data, and Knowledge Graphs workshop (VOILA!). The paper describes an approach for navigating the knowledge of privacy engineering. Co-authored with Fajar J. Ekaputra, Florina Piroi, and Marta Sabou, the paper emphasizes the need for efficient navigation of interlinked information in privacy engineering, especially in privacy threat modeling.
The proposed solution is an exploratory search system developed by the semantic systems group in the OntoTrans project. This system allows stakeholders to search the privacy knowledge such as threats, relationships with the data flow diagrams, and identify new protective measures against privacy risks. By supporting learning and exploratory tasks, this interactive model addresses the dynamic challenges of privacy engineering and provides a proactive approach to privacy protection.
For more insights into the exploratory search system, visit the documentation.
Conference Highlights
Day 1: Workshops and Tutorials
The conference kicked off with a series of workshops and tutorials. A notable session was the Bosch Tutorial on Neuro-Symbolic AI, which explored various approaches to semantic data integration. The tutorial highlighted the increasing interest from industries like Siemens and Bosch in semantic AI technologies, particularly for applications like semantic digital twins and industrial vision.
Day 2: Doctoral Symposium and Workshops
The second day featured a doctoral symposium where emerging researchers presented their work on topics ranging from semantic-based metaphors in research articles to semantic web and NLP for process identification in the different domains. The day also included discussions on the integration of knowledge graphs with large language models, highlighting the challenges and opportunities in this area.
Day 3: Main Conference
The main conference began with a keynote on “Knowledge Graphs in the Age of Large Language Models” by Gerhard Weikum. The talk delved into the strengths and limitations of knowledge graphs, search engines, and language models, and proposed a three-stage architecture integrating these components for more effective information retrieval and understanding.
Further Insights
The final days continued with more insightful keynotes and sessions. Deborah McGuinness’s keynote on “Semantic Web Research in the Age of Generative Artificial Intelligence” was particularly thought-provoking, discussing the potential renaissance in semantic web research and methodologies. The conference also showcased various industry applications, highlighting the practical impact of semantic web technologies.
Conclusion
The ISWC conference was an enlightening experience, filled with learning and networking opportunities. It was inspiring to see the diverse applications of semantic web technologies and the innovative research being conducted in this field. Our team’s participation and contributions to the conference have been immensely rewarding, and we look forward to applying these insights to our future projects.
References:
X. Li, A. Hughes, M. Llugiqi, F. Polat, P. Groth, & F.J Ekaputra. Knowledge-centric Prompt Composition for Knowledge Base Construction from Pre-trained Language Models, LM-KBC, 2023.
G. B. Herwanto, F.J. Ekaputra, F. Piroi, & M. Sabou. Towards A Knowledge Graph-based Exploratory Search for Privacy Engineering, VOILA, 2023.
