Efficient Reinforcement Learning with Multiple Reward Functions for Randomized Controlled Trial Analysis

By Daniel J. Lizotte, Michael H. Bowling, Susan A. Murphy. 2010

88 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Informing sequential clinical decision-making through reinforcement learning: an empirical study (2010, background): “problem (Bagnell et al. 2001; Shelton 2001; Doshi et al. 2008; Lizotte et al. 2010) but more”
Multi-objective reinforcement learning using sets of pareto dominating policies (2014, methodology): “Lizotte et al. (2010) reduce the asymptotic space and time complexity of the bootstrapping rule by learning several value functions corresponding to different weight vectors using a piecewise linear spline representation.”
A Survey of Multi-Objective Sequential Decision-Making (2013, methodology): “To our knowledge, the only work in this area is that of Lizotte et al. (2010, 2012), where a model of the MOMDP’s transition probabilities and reward function is derived from historical data, and then a spline-based multi-objective value iteration approach is applied to that model.”
Mobile Health Technology Evaluation (2013, methodology): “In SMART, researchers decide which aspects of treatments require investigation and then randomize individuals at each treatment decision based on feasibility, ethical issues, or other factors.”
Linear Fitted-Q Iteration with Multiple Reward Functions (2012, methodology): “Lemma 5 (Lizotte et al., 2010) If action a is optimal at time T for some point (ψ s T , δ) but is not optimal for any (ψ s T , δ) on the boundary of the domain, then a is optimal for some (ψ s T , δ) that is a triple-point.”
Scalarized multi-objective reinforcement learning: Novel design techniques (2013, background): “Additionally, [6] also proposes a batch MORL approach, based on the linear scalarization function, to identify which actions are favoured in which parts of the objective space.”