neXSim

About This Project

Nowadays, discovering new entities similar to ones already aligned with users' preferences is a common and challenging task. However, typically, users are not inclined to express their tastes explicitly, so extracting such information as similarities among entities they already approved is crucial. Various recommendation systems move in this direction, but relying on machine learning and statistical techniques, they lack in supporting their output via a user-readable explanation. Accordingly, recent research introduced a theoretical deductive logic-based framework to formally express similarity among a set of entities (tuples of entities), compare and expand the latter with others sharing the same nexus of similarity or a more generic one. This work proposes a practical application of this framework, proving its feasibility by implementing RESTful APIs integrated into a web-based system to characterize similarities among entities and, later, recommend new ones according to their nexus of similarity while providing a formal and well-formed explanation. For this purpose, the system draws upon a knowledge base with data extracted from a prior version of the well-known BabelNet knowledge graph. Yet, its intrinsic design allows it to adapt and operate over a generic knowledge base.

Scientific References

A logic-based framework for characterizing nexus of similarity within knowledge bases.
G. Amendola, M. Manna, A. Ricioppo
View Source
Cite this article
Characterizing Nexus of Similarity within Knowledge Bases: A Logic-based Framework and its Computational Complexity Aspects.
G. Amendola, M. Manna, A. Ricioppo
View Source
Cite this article
Characterizing Nexus of Similarity between Entities.
G. Agresta, G. Amendola, P. Cofone, M. Manna, A. Ricioppo
View Source
Cite this article

Acknowledgments

In this version of neXSim, the underlying Knowledge Base relies on BabelNet 4.0.1
This work was funded by:
  • FAIR – “Future AI Research” (CUP: PE00000013), Spoke 9 – Green-aware AI, under the NextGenerationEU NRRP MUR program.
  • Tech4You – “Technologies for climate change adaptation and quality of life improvement” (CUP: ECS0000009), Spoke 6 – ICT for Digital Transformation, under the NextGenerationEU NRRP MUR program.
  • PRODE – “Probabilistic declarative process mining” (CUP: H53D23003420006), under the PRIN MUR program.