Integrating Top-Down Planning with a Bottom-Up Approach that Learns to Locally Combine Actions.
Blazej BulkaMarie desJardinsPublished in: AAAI Spring Symposium: Distributed Plan and Schedule Management (2006)
Keyphrases
- stochastic domains
- decision theoretic
- plan recognition
- macro actions
- attention mechanism
- goal directed
- action selection
- classical planning
- world model
- external events
- plan generation
- plan execution
- high level
- concurrent actions
- visual saliency
- partially observable
- intelligent behavior
- partial plans
- action sequences
- visual attention
- goal state
- initial state
- learning algorithm
- reinforcement learning
- state space
- heuristic search
- figure ground segmentation
- sensing actions
- domain independent
- decision theoretic planning
- planning graph
- ai planning
- figure ground
- markov decision processes
- state action
- situation calculus
- reasoning about actions
- human activities
- planning process