Quote taken from the last paragraph of section 4 of 'A Generative Inquiry Dialogue System' by Elizabeth Black and Anthony Hunter
In our simple example, it seems that it would be more straightforward to pool both agents beliefs and apply a reasoning procedure to this set of beliefs. However, given a real-world scenario this would not necessarily be the case. When dealing with the medical domain we have to consider privacy issues that would restrict agents from simply pooling all beliefs. It is also sometimes the case that agents have vast belief bases and the communication cost involved in sharing all beliefs would be prohibitive.
Post a Comment