The paper gives the first tight necessity and sufficiency conditions for successful reward poisoning attacks in linear MDPs.
Stage 1 certifies by time T1 and yields a fixed attacked MDP on which the target policy is best on the certified support
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
When Can You Poison Rewards? A Tight Characterization of Reward Poisoning in Linear MDPs
The paper gives the first tight necessity and sufficiency conditions for successful reward poisoning attacks in linear MDPs.