# A Critical Analysis of Recursive Model Indexes

arXiv: https://arxiv.org/abs/2106.16166 · HTML: https://mlsanity.com/paper/2106.16166/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Jun 2021.

## Built on by (most cited first, 9 of 9)

- [Learned Sorted Table Search and Static Indexes in Small Model Space](https://mlsanity.com/paper/2107.09480/) (2021)
- [On the Choice of General Purpose Classifiers in Learned Bloom Filters: An Initial Analysis Within Basic Filters](https://mlsanity.com/paper/2112.06563/) (2021)
- [A Survey of Learned Indexes for the Multi-dimensional Space](https://mlsanity.com/paper/2403.06456/) (2024)
- [Standard Vs Uniform Binary Search and Their Variants in Learned Static Indexing: The Case of the Searching on Sorted Data Benchmarking Software Platform](https://mlsanity.com/paper/2201.01554/) (2022)
- [The Duck's Brain: Training and Inference of Neural Networks in Modern Database Engines](https://mlsanity.com/paper/2312.17355/) (2023)
- [LSI: A Learned Secondary Index Structure](https://mlsanity.com/paper/2205.05769/) (2022)
- [A Critical Analysis of Classifier Selection in Learned Bloom Filters](https://mlsanity.com/paper/2211.15565/) (2022)
- [LearnedSort as a learning-augmented SampleSort: Analysis and Parallelization](https://mlsanity.com/paper/2307.08637/) (2023)
- [From Specific to Generic Learned Sorted Set Dictionaries: A Theoretically Sound Paradigm Yelding Competitive Data Structural Boosters in Practice](https://mlsanity.com/paper/2309.00946/) (2023)

## Abstract (arXiv)

The recursive model index (RMI) has recently been introduced as a machine-learned replacement for traditional indexes over sorted data, achieving remarkably fast lookups. Follow-up work focused on explaining RMI's performance and automatically configuring RMIs through enumeration. Unfortunately, configuring RMIs involves setting several hyperparameters, the enumeration of which is often too time-consuming in practice. Therefore, in this work, we conduct the first inventor-independent broad analysis of RMIs with the goal of understanding the impact of each hyperparameter on performance. In particular, we show that in addition to model types and layer size, error bounds and search algorithms must be considered to achieve the best possible performance. Based on our findings, we develop a simple-to-follow guideline for configuring RMIs. We evaluate our guideline by comparing the resulting RMIs with a number of state-of-the-art indexes, both learned and traditional. We show that our simple guideline is sufficient to achieve competitive performance with other learned indexes and RMIs whose configuration was determined using an expensive enumeration procedure. In addition, while carefully reimplementing RMIs, we are able to improve the build time by 2.5x to 6.3x.