# Learning to select computations

arXiv: https://arxiv.org/abs/1711.06892 · HTML: https://mlsanity.com/paper/1711.06892/

## Key facts

- Cited by 15 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Nov 2017.

## Built on by (most cited first, 15 of 15)

- [When Can Model-Free Reinforcement Learning be Enough for Thinking?](https://mlsanity.com/paper/2506.17124/) (2025)
- [Improving Human Decision-Making by Discovering Efficient Strategies for Hierarchical Planning](https://mlsanity.com/paper/2102.00521/) (2021)
- [Boosting human decision-making with AI-generated decision aids](https://mlsanity.com/paper/2203.02776/) (2022)
- [Toward Agents That Reason About Their Computation](https://mlsanity.com/paper/2510.22833/) (2025)
- [Have I done enough planning or should I plan more?](https://mlsanity.com/paper/2201.00764/) (2022)
- [What are the mechanisms underlying metacognitive learning?](https://mlsanity.com/paper/2302.04840/) (2023)
- [Finding the Time to Think: Learning Planning Budgets in Real-Time RL](https://mlsanity.com/paper/2606.26463/) (2026)
- [Modeling Human Behavior Part I -- Learning and Belief Approaches](https://mlsanity.com/paper/2205.06485/) (2022)
- [An intelligent tutor for planning in large partially observable environments](https://mlsanity.com/paper/2302.02785/) (2023)
- [Experience-driven discovery of planning strategies](https://mlsanity.com/paper/2412.03111/) (2024)
- [Artificial Agency Program: Curiosity, compression, and communication in agents](https://mlsanity.com/paper/2602.24100/) (2026)
- [When to Plan: Learning to Select Between Reactive Control and Deliberative Planning](https://mlsanity.com/paper/2607.16421/) (2026)
- [Discovering the curriculum with AI: A proof-of-concept demonstration with an intelligent tutoring system for teaching project selection](https://mlsanity.com/paper/2406.04082/) (2024)
- [Individual differences in the cognitive mechanisms of planning strategy discovery](https://mlsanity.com/paper/2505.23519/) (2025)
- [A Vision Toward Energy-Efficient Domain-Specific Artificial Intelligence Models and Agents](https://mlsanity.com/paper/2510.22052/) (2025)

## Abstract (arXiv)

The efficient use of limited computational resources is an essential ingredient of intelligence. Selecting computations optimally according to rational metareasoning would achieve this, but this is computationally intractable. Inspired by psychology and neuroscience, we propose the first concrete and domain-general learning algorithm for approximating the optimal selection of computations: Bayesian metalevel policy search (BMPS). We derive this general, sample-efficient search algorithm for a computation-selecting metalevel policy based on the insight that the value of information lies between the myopic value of information and the value of perfect information. We evaluate BMPS on three increasingly difficult metareasoning problems: when to terminate computation, how to allocate computation between competing options, and planning. Across all three domains, BMPS achieved near-optimal performance and compared favorably to previously proposed metareasoning heuristics. Finally, we demonstrate the practical utility of BMPS in an emergency management scenario, even accounting for the overhead of metareasoning.