paper
Action-BED: Task-Driven Bayesian Experimental Design with Singly Intractable Objectives
paperactiveprovisional
action-bed-task-driven-bayesian-experimental-design-with-singly-intractable-objectives-dbb8847e·1 events·first seen 5d agoAliases: Action-BED: Task-Driven Bayesian Experimental Design with Singly Intractable Objectives
Co-occurring entities
More like this (12)
Bayesian Optimal Experimental DesignACTION-BEDBayesian OptimizationBayesian Multiobjective OptimizationAction-Dependent Factorized BaselinesAPPO: Agentic Procedural Policy Optimizationagent-task efficiencytask-conditioned attractorsMatch Task to Objective (MTO)UniIntervene: Agentic Intervention for Efficient Real-World Reinforcement LearningGeometric Action Model for Robot Policy LearningBehavioral Trajectory Tracking Framework
Recent events (1)
ACTION-BED: Task-driven Bayesian experimental design via singly intractable expected future loss objectives
A new arXiv preprint proposes ACTION-BED, a reformulation of Bayesian experimental design (BED) that replaces the traditional doubly intractable expected information gain objective with an expected future loss (EFL) on downstream actions. The authors show all such EFLs can be rearranged into singly intractable objectives jointly optimizable over design and action policies via stochastic gradients, eliminating the need for explicit posterior or marginal likelihood estimation. The method is claimed to be more efficient and customizable to downstream tasks than existing BED approaches.