Distributed AI in the Modern World: Technical and Social Aspects of Interacting Intelligent Agents
presents several state-of-the-art insights into the various forms of distribution of artificial intelligence, with practical application instances.
This book does not analyze the internal workings of machine learning models (for instance, in the case of multi-agent reinforcement learning), but instead provides readers with an overview of the challenges brought by the need of artificially intelligent entities to interact with other entities and with their environment along with practical solutions at an architectural level.
Deployment, maintenance and monitoring of distributed machine learning systems brings about many practical challenges, dealing with the intelligent agents distributed across a network of heterogenous devices, or interacting with robots and humans alike.
While these scenarios are very different, some challenges remain the same when interaction exists: discoverability, availability, communication language and formats, and efficiency in transferring significant amounts of information.
The book provides readers with practical solutions at an architectural level, with solutions presented in three parts.
Part 1 deals with the distribution of the learning process and the utilization of machine learning models in a distributed system.
Part 2 deals with tools that enable the distribution and interaction of artificial learning entities and how multi-agent systems and machine learning can be combined.
Part 3 deals with the physical embodiment of intelligent agents and the interaction of intelligent computing units bound to physical space.
The three parts are followed by a conclusion, emphasizing the challenges that are common to all scenarios and solutions which apply in a wider range of cases.
Reviews
No Review Found