Explainable AI and logic-based Intent management and heuristics and DRL for resource orchestration

Explainable AI and logic-based Intent management and heuristics and DRL for resource orchestration

Our research expertise is in the area of network and service resource orchestration and zero-touch closed-loop management in the edge cloud continuum and b5g networks with focus on:
- New design approaches for 6G system architecture systems
- Native and trustworthy integration of AI for telecommunications

In particular expertise is offered in the following areas:
- ***Explainable AI and logic-based approaches for Intent management***, including: intent processing, translation, conflict detection and solution, negotiation with users. Intent is a key enabler for zero-touch automation. Intent only defines the goals and expectations, therefore translation and processing mechanisms are required to manage intents across multiple layers (vertical layer, network service layer, infrastructure layer) considering possible conflicts among stakeholders, such those among verticals’ requirements and SLOs and infrastructure provider long term expectations in terms of sustainability, energy efficiency and trustworthiness, etc. We developed an open source intent processing and conflict detection and resolution prototype.
- ***Reinforcement learning strategies for resource configuration orchestration and allocation*** and AI/ML-driven intent translation across layers and across technology domains. We developed an open source Deep Reinforcement Learning toolkit for Network Slicing.
- Design, implementation and experimental analysis of ***resource configuration orchestration and allocation algorithms***. We developed several solution strategies, including heuristics, matheuristics and metaheuristics.

not-set