Algolog Master thesis defence Friday 3 July
8 views
Skip to first unread message
Thomas Bolander
Jun 30, 2015, 8:06:36 AM6/30/15
to <algolog@googlegroups.com>, COMP-all_algolog, Mikko Berggren Ettienne
Friday 3 July at 13.00, Søren Jacobsen and Jannick Johnsen will defend their master thesis with the title
"Action Learning in Automated Planning".
The defence will take place in S02 in the meeting centre in building 101. All are welcome.
Best,
Thomas
--
Summary:
This thesis attempts to solve the problem of action schema learning. That is, learning actions in a domain without knowledge of those actions’ effects and precondition, and only by analyzing the state-transitions. The solution to these problems should not use any probability or evolutionary theories, but rather infer the actions, based only on analyzing the state-transitions. This thesis extends the work of Walsh and Littman [2008], by exploring the underlying properties of their work. Because of this we are able to remove some of the restrictions in their model.
The goal of this thesis is to elucidate what it means to learn. Specifically we will focus on non-conditional actions, and attempt to provide a framework to solve the problem of conditional action schema learning.
We solve these problems using graph theory, set theory and logic reasoning.
"Action Learning in Automated Planning".
The defence will take place in S02 in the meeting centre in building 101. All are welcome.
Best,
Thomas
--
Summary:
This thesis attempts to solve the problem of action schema learning. That is, learning actions in a domain without knowledge of those actions’ effects and precondition, and only by analyzing the state-transitions. The solution to these problems should not use any probability or evolutionary theories, but rather infer the actions, based only on analyzing the state-transitions. This thesis extends the work of Walsh and Littman [2008], by exploring the underlying properties of their work. Because of this we are able to remove some of the restrictions in their model.
The goal of this thesis is to elucidate what it means to learn. Specifically we will focus on non-conditional actions, and attempt to provide a framework to solve the problem of conditional action schema learning.
We solve these problems using graph theory, set theory and logic reasoning.
Reply all
Reply to author
Forward
0 new messages