Logical Agents: Memory Management, Advanced Architectures and Applications

By Valentina Pitoni,
Dipartimento di Ingegneria e Scienze dell’Informazione e Matematica, Università degli Studi dell’Aquila
2020

Abstract

In this thesis I show my work of the past 3 years. I have covered different topics and the main areas are:

  • Modal, Temporal and Metric Temporal Logic and Epistemic Logic;
  • Cyber-Physical Systems, Multi-Context Systems, Component-based Agents Environments;
  • Machine Ethics.

The primary objective of my research has been the study of autonomous system which evolves over time and their formal treatment in Computational Logic. First of all I show how I manage agents’ memory through a particular modal logic. I have started to study this topic from the Master thesis and I carried on the work introducing various extensions that I will describe in the first chapter.

In the second chapter I show my last work which is a “work in progress”. It is about a particular logic which is similar to the one explained in the first chapter, but I underline the concepts of “steps” and “executability’”. These two characterize agents’ action.

In the third chapter I show my other research areas: Cyber-Physical Systems. Here I focus on work about a particular, innovative kind of architecture called K-Level ACE, which has a fractal structure and than can manage sudden failures.

In the forth chapter I show the last topic of my research activities: Machine Ethics, which is a part of ethics of Artificial Intelligence concerned with the moral behavior of artificial intelligent beings. Important aspect of Machine Ethics are trustworthiness and safety. We accomplished this aspects through verification and assurance. We propose technique for Runtime self checking and Monitoring using meta-rules and runtime constraints.

FULL TEXT