Rashid Ayub
Research Scholar, Sunrise University, Alwar
Prasadu Peddi
Professor, Sunrise University, Alwar
Abstract
The apparatus we used to test and refine our Bounded Rationality HMM answer to the BRAI. In particular, we looked at two simulation settings, each of which has its own quirks. To begin, we used Deep learning, a fully visible robotic mining simulation in which an agent must use a microscope to detect the mineral composition of a mine, and then gather those minerals in order to perform tasks. The microscope's accuracy degrades during operation (depending on the sort of test being conducted), but recovers over time when it is not in use. Second, we used a simulation for partly observable user preference elicitation dubbed User Rec, which is based on the work of Doshi and Roy (2008) and describes a scenario in which an intelligent user interface agent must ascertain a user's choice through interruptions that solicit data from the user.
Keywords: Deep Learning, Recognition, Sensor-Based, Human Activity