The bayesian approach to belief propagation in digital ecosystems
A. Tang · 2009
Bayesian belief propagation is flexible and highly adaptable in not only machine learning and artificial intelligence methodologies, but also to newer forms of learning involving agent interactions in digital ecosystems, specifically multi-agent systems. One important property of such systems is agent autonomy. An aspect of agent autonomy, enactive knowledge, is investigated here through a Bayesian extension called TAN that supports learning through interactions with the environment. Finally, various scenarios are simulated for an appropriate modelling environment with suggestions for future work.