AT-UCB uses a cheap abstracted causal model to filter out suboptimal actions before running UCB on the expensive base model, with a regret bound that improves when the abstraction is accurate.
Causal bandits: Learning good interventions via causal inference
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Using causal abstractions to accelerate decision-making in complex bandit problems
AT-UCB uses a cheap abstracted causal model to filter out suboptimal actions before running UCB on the expensive base model, with a regret bound that improves when the abstraction is accurate.