Efficient Learning of Action Models for Planning
نویسندگان
چکیده
We consider the problem of learning action models for planning in two frameworks and present general sufficient conditions for efficient learning. In the mistake-bounded planning framework, the learner has access to a sound and complete planner for the given action model language, a simulator, and a planning problem generator. In the planned exploration framework, the learner has access to a planner and a simulator, but actively generates problems to help refine its model. We identify sufficient conditions for learning in both the frameworks. We also show that a concrete hypothesis space that consists of sets of rules with at most k variables is efficiently learnable in both frameworks.
منابع مشابه
Active model learning and diverse action sampling for task and motion planning
The objective of this work is to augment the basic abilities of a robot by learning to use new sensorimotor primitives to enable the solution of complex long-horizon problems. Solving long-horizon problems in complex domains requires flexible generative planning that can combine primitive abilities in novel combinations to solve problems as they arise in the world. In order to plan to combine p...
متن کاملThe Impact of Studio-based learning on Metacognition and Design Ability of Architecture Students - Action Research
Proper training can put design learners in the right direction. It also enhances the power of drawing. Objective of this study was the effectiveness of architectural studio-based learning on increasing drawing power and metacognition abilities of students. This research seeks to answer these questions: Can architectural studio-based learning increase student design ability? Can architectural st...
متن کاملBehavioural priors: Learning to search efficiently in action planning
We propose that an important part of planning (modelbased reinforcement learning) in a biological agent is to use information on the statistical structure of nearoptimal action sequences. Such statistical information can be collected simply by observing the action sequences that the agent has actually performed, assuming that the agent only performs action sequences that are the most optimal on...
متن کاملFrom Non-Deterministic to Probabilistic Planning with the help of Statistical Relational Learning
Using machine learning techniques for planning is getting increasingly more important in recent years. Various aspects of action models can be induced from data and then exploited for planning. For probabilistic planning, natural candidates are learning of action effects and their probabilities. For expressive formalisms such as PPDDL, this is a difficult problem since they can introduce easily...
متن کاملLearning HTN Method Preconditions and Action Models from Partial Observations
To apply hierarchical task network (HTN) planning to real-world planning problems, one needs to encode the HTN schemata and action models beforehand. However, acquiring such domain knowledge is difficult and time-consuming because the HTN domain definition involves a significant knowledge-engineering effort. A system that can learn the HTN planning domain knowledge automatically would save time...
متن کامل