SymNet 2.0: Effectively handling Non-Fluents and Actions in Generalized Neural Policies for RDDL Relational MDPs.
Vishal SharmaDaman AroraFlorian Geißer MausamParag SinglaPublished in: UAI (2022)
Keyphrases
- reward function
- situation calculus
- optimal policy
- markov decision processes
- initial state
- decision processes
- concurrent actions
- reasoning about actions
- decision theoretic planning
- action language
- partially observable
- temporally extended
- markov decision process
- markov decision problems
- reinforcement learning
- state space
- planning problems
- relational data
- macro actions
- state and action spaces
- data model
- decision diagrams
- stochastic domains
- policy search
- action space
- indirect effects
- fitted q iteration
- network architecture
- neural network
- multiagent reinforcement learning
- decision theoretic
- relational databases
- reinforcement learning algorithms
- discounted reward
- finite horizon
- policy iteration
- average cost
- action theories
- atomic actions
- multiple agents
- event calculus
- control policies