Abstract

Integrating Learning and Reasoning is a fundamental problem in AI, especially in application domains dealing with relational data, such as knowledge graphs and ontologies. In particular, the notion of embedding may play a crucial role to encode relational knowledge in a latent space and to provide a more flexible representation to perform learning and reasoning.

During this workshop, some relevant models and methods employing different kind of reasoning mechanisms will be presented and discussed. In particular, the goal will be to present some of the main research activities of the CINI group in order to outline possible collaboration and research directions around Task 4.3. Moreover, thanks to the partecipation of Task 7.4 of AI4EU project, it will be discussed how newly developed assets for the integration of learning and reasoning might be published on the AI4EU platform.

Program

Program

• 09:30-09:40 - Welcome and Introduction (Marco Gori, UniSi)

• 09:40-10:25 - First Talk Session (Chair Marco Lippi)

  • KENN: Knowledge Enhanced Neural Networks (Alessandro Daniele, FBK)

  • Learning Representation for Sub-Symbolic Reasoning (Francesco Giannini, UniSi)

  • Empirical Model Learning: embedding ML models in declarative optimization model (Michele Lombardi, UniBo)

• 10:25-10:50 - How you can publish your work on the AI4EU platform (Alessandro Saffiotti, ORU & Peter Schuller, TUW)

• 10:50-11:05 - Coffee break

• 11:05-11:50 - Second Talk Session (Chair Francesco Giannini)

  • Structure Learning of Probabilistic Logic Programs (Fabrizio Riguzzi, UniFe)

  • Towards Explainable Autonomous Development (Marco Lippi, UniMoRe)

  • Online Learning of Planning Domain Representations from Sensor Data (Alfonso Gerevini & Leonardo Lamanna, UniBs)

• 11:50-12:25 - Open issues & Proposals around the task

• 12:25-13:00 - Follow-up discussion & Collaboration definition