Semantic Web: Best Paper Award for Yiwen Peng, Thomas Bonald, and Fabian Suchanek

16 February 2026

Fabian Suchanek, Yiwen Peng, Thomas Bonald

Yiwen Peng, a PhD candidate, together with Thomas Bonald and Fabian Suchanek, researchers at Télécom Paris – Institut Polytechnique de Paris, have been awarded the Best Paper Award at the 24^th International Semantic Web Conference (ISWC 2025) for their paper entitled “FLORA: Unsupervised Knowledge Graph Alignment by Fuzzy Logic”, published in the conference proceedings.

International Semantic Web Conference ISWC is the leading international conference in the field of the Semantic Web and knowledge graphs, annually bringing together cutting-edge scientific contributions on semantic technologies, structured knowledge, and symbolic artificial intelligence. The Best Paper Award recognizes contributions that the program committee considers particularly innovative and promising.

At the intersection of knowledge graphs and fuzzy logic

The award-winning paper introduces FLORA, a novel method for knowledge graph alignment, a core task in data science and artificial intelligence. Knowledge graph alignment consists in identifying and associating equivalent entities and relations across two graphs—for example, between databases such as Wikidata and YAGO—in order to merge information from heterogeneous sources and thereby enhance knowledge interoperability.

This problem is fundamental to numerous application domains, including:

the integration of data from diverse sources,
answering complex queries over multiple knowledge bases, and
symbolic reasoning in artificial intelligence.

Traditional alignment approaches often rely on supervised learning techniques or vector representations (embeddings) that require annotated training data, which limits their practical applicability.

An interpretable, unsupervised, and robust approach

FLORA is distinguished by its fully unsupervised design: the method requires no annotated training data, making it suitable for scenarios in which such resources are scarce or costly to obtain. It is grounded in principles of fuzzy logic, providing an interpretable framework for decision-making processes. This approach enables the effective alignment not only of entities but also of relations, while offering mathematical guarantees of convergence and the ability to handle dangling entities (i.e., entities without counterparts).

The experimental results reported in the paper demonstrate that FLORA achieves state-of-the-art performance on several established benchmarks, confirming both its scientific relevance and its practical value.

Strengthening Télécom Paris’ scientific visibility

This distinction highlights the excellence of research conducted at Télécom Paris in the areas of the Semantic Web, knowledge graphs, and artificial intelligence. It underscores the ability of the institution’s research teams to produce high-impact work and to make significant contributions to advances within the international scientific community.

We extend our warmest congratulations to Yiwen Peng, Thomas Bonald, and Fabian M. Suchanek (Data, Intelligence, and Graphs team at Télécom Paris’ LTCI lab) on this outstanding achievement.