Bi-Weekly Talk: Daxin Liu: Reasoning about uncertainty

Wednesday, June 12, 2019, 10:30am

Location: RWTH Aachen University, Department of Computer Science - Ahornstr. 55, building E3, room 9u10

Speaker: Daxin Liu

 

abstract: 

When modeling agents' belief in a dynamic uncertain domain, a major concern is how to better relate high-level belief with low-level actions' likelihood. This enforces the modeler must take the actions' observational-indistinguishability into account. Meanwhile, since noise sensing is the only way for the agent to increase knowledge or enhance belief, suitable semantic needs to be assigned to it. In this talk, we first introduce a model logic DS* that capture all above, then explore a projection reasoning mechanism in such logic, namely regression.