Adversarial poisoning attacks on reinforcement learning-driven energy pricing.
Sam GunnDoseok JangOrr ParadiseLucas SpangherCostas J. SpanosPublished in: BuildSys@SenSys (2022)
Keyphrases
- reinforcement learning
- multi agent
- energy consumption
- function approximation
- data driven
- state space
- machine learning
- markov decision processes
- energy minimization
- countermeasures
- optimal policy
- learning process
- watermarking scheme
- reinforcement learning algorithms
- model free
- malicious users
- profit maximization
- malicious attacks
- security threats
- dynamic pricing
- learning algorithm
- security mechanisms
- terrorist attacks
- dynamic programming