Myopic Bayesian Design of Experiments via Posterior Sampling and Probabilistic Programming
arXiv:1805.09964
Abstract
We design a new myopic strategy for a wide class of sequential design of experiment (DOE) problems, where the goal is to collect data in order to to fulfil a certain problem specific goal. Our approach, Myopic Posterior Sampling (MPS), is inspired by the classical posterior (Thompson) sampling algorithm for multi-armed bandits and leverages the flexibility of probabilistic programming and approximate Bayesian inference to address a broad set of problems. Empirically, this general-purpose strategy is competitive with more specialised methods in a wide array of DOE tasks, and more importantly, enables addressing complex DOE goals where no existing method seems applicable. On the theoretical side, we leverage ideas from adaptive submodularity and reinforcement learning to derive conditions under which MPS achieves sublinear regret against natural benchmark policies.
References in corpus (8)
- Predictive Entropy Search for Multi-objective Bayesian Optimization
- Reinforcement and Imitation Learning via Interactive No-Regret Learning
- Deep Probabilistic Programming
- Fast Variational Inference in the Conjugate Exponential Family
- Parallel and Distributed Thompson Sampling for Large-scale Accelerated Exploration of Chemical Space
- Near-optimal Reinforcement Learning in Factored MDPs
- Asynchronous Parallel Bayesian Optimisation via Thompson Sampling
- Embarrassingly Parallel Variational Inference in Nonconjugate Models