As part of the Canadian Distributed Mentorship Program, here is where I will be posting my journal entries relating to the work done.

I -

II -

III -

IV -

V -

VI -

VII -

VIII -

IX-

X-

XI-

XII-

XIII-

XIV-

XV-

XVI-

XVII-

XVIII

I've continued to work on the Cat and Mouse environment, and progress can finally be seen! Not only does the environment compile, it allows an agent to connect to it without breaking networking havoc. As such, I've developed two learning agents, in both C and Java to test this interaction. 'Learning agent' is an overstatement, however, as the agents perform entirely random actions, and the poor mouse often finds himself as chef's special.

The general idea is as follows: the agent receives its current state (the cat, mouse and cheese's position) from the environment. Based on this state, it performs an action, and sends it back to the environment; the environment, in turn, executes the action, observes the new state that results from it, and sends a reward back to the agent.
The final version of the Cat and Mouse simulator can be found here.

This week Doina, Eric and I had the first meeting to discuss the Nortel project. The first stage has to do with inferring about daily actions of a doctor or nurse in a hospital. Each doctor/nurse/surgeon etc. is to wear a tracking device, that would gather movement and action information. By developing a set of policies, the idea is to choose an appropriate emergency response team if an emergency event is to occur in the hospital.
This all sounds much more confusing than it is. Take Bill and Bob, surgeons with fancy tracking devices, and Jane and Julia nurses. Bob is currently performing complicated brain surgery on Julia, and Bill is having a possibly moreso complicated double-mocha-latte in the staff room with Jane. Armed with information about the fantastic four, as well as the news that Joe is being rushed to the hospital the application ought to send Bill and Jane to his aid, as they are the both available and qualified. Of course, the challenge is to develop the policy according to which actions should be taken, based on a much larger set of states, persons and roles...