About the Project
KNOW-NET-COMPET – Embedded Knowledge and Networks: Complexity in Antitrust Collusion Mitigation
Coordinated by NOVA IMS, the KNOW-NET-COMPET project responds to the increasing demand for innovative tools to detect and prevent collusion and anti-competitive practices in modern markets. The initiative is pioneering the creation of an interpretable, data-driven decision-support system grounded in network science and knowledge representation.
The project leverages advances in Natural Language Processing, Knowledge Graphs, and Graph Representation Learning to map complex market structures and uncover hidden links indicative of collusion. It aligns with the emerging field of computational antitrust, responding to the digitalization of markets and regulatory demands for transparent, AI-supported enforcement mechanisms.
Through three interlinked phases—data cataloging and review, knowledge graph development, and analytical validation—the project produces a prototype platform that integrates structured and unstructured data sources, while advancing state-of-the-art research in explainable AI and graph analytics.
As coordinator, NOVA IMS applies its advanced expertise in data science, AI, and economic complexity to:
- Design and implement a Knowledge Graph integrating legal, institutional, and market data to model competition-relevant networks.
- Develop analytical modules to predict collusion risks via link prediction and market segmentation using graph embedding methods.
- Create a dashboard interface to explore indicators and interact with the graph in real-time, facilitating decision-making and regulatory interventions.
- Produce open-access scientific publications, software prototypes, and a public final seminar, promoting transparency and academic impact.
