Causally Abstracted Multi-armed Bandits
This work addresses the need for joint formulations in decision-making for related problems with multi-scale observations, representing an incremental extension of transfer learning for causal MABs.
The paper tackles the problem of decision-making across multiple related causal multi-armed bandit problems with different variables and granularities by introducing causally abstracted MABs (CAMABs) and proposing algorithms to learn in this setup, achieving results with studied regret bounds and illustrated on a real-world online advertising scenario.
Multi-armed bandits (MAB) and causal MABs (CMAB) are established frameworks for decision-making problems. The majority of prior work typically studies and solves individual MAB and CMAB in isolation for a given problem and associated data. However, decision-makers are often faced with multiple related problems and multi-scale observations where joint formulations are needed in order to efficiently exploit the problem structures and data dependencies. Transfer learning for CMABs addresses the situation where models are defined on identical variables, although causal connections may differ. In this work, we extend transfer learning to setups involving CMABs defined on potentially different variables, with varying degrees of granularity, and related via an abstraction map. Formally, we introduce the problem of causally abstracted MABs (CAMABs) by relying on the theory of causal abstraction in order to express a rigorous abstraction map. We propose algorithms to learn in a CAMAB, and study their regret. We illustrate the limitations and the strengths of our algorithms on a real-world scenario related to online advertising.