8 Hits in 1.5 sec

RadixSpline: A Single-Pass Learned Index [article]

Andreas Kipf, Ryan Marcus, Alexander van Renen, Mihail Stoian, Alfons Kemper, Tim Kraska, Thomas Neumann
2020 arXiv   pre-print
We introduce RadixSpline (RS), a learned index that can be built in a single pass over the data and is competitive with state-of-the-art learned index models, like RMI, in size and lookup performance.  ...  Recent research has shown that learned models can outperform state-of-the-art index structures in size and lookup performance.  ...  CONCLUSIONS We have described a new learned index, called RadixSpline, that can be built in a single pass over sorted data.  ... 
arXiv:2004.14541v2 fatcat:wrxikhau2ve47df64nio2ecs2i

A Critical Analysis of Recursive Model Indexes [article]

Marcel Maltry, Jens Dittrich
2021 arXiv   pre-print
The recursive model index (RMI) has recently been introduced as a machine-learned replacement for traditional indexes over sorted data, achieving remarkably fast lookups.  ...  We evaluate our guideline by comparing the resulting RMIs with a number of state-of-the-art indexes, both learned and traditional.  ...  The linear spline is fit in a single pass over the data and guarantees a user-defined error bound.  ... 
arXiv:2106.16166v2 fatcat:lmrp445dhrgfxaczxhx2o2h4pa

The Case for Learned Spatial Indexes [article]

Varun Pandey, Alexander van Renen, Andreas Kipf, Ibrahim Sabek, Jialin Ding, Alfons Kemper
2020 arXiv   pre-print
learned indexes is 1.23x to 1.83x times faster than closest competitor which filters on two dimensions, and (iv) learned indexes can have a significant impact on the performance of low selectivity queries  ...  In this work, we use techniques proposed from a state-of-the art learned multi-dimensional index structure (namely, Flood) and apply them to five classical multi-dimensional indexes to be able to answer  ...  Unlike the RMI [23] , the RadixSpline only requires one pass over the data to build the index, while retaining competitive lookup times.  ... 
arXiv:2008.10349v1 fatcat:slbhet4gdzh55ilm6j66qv6gjy

Towards Practical Learned Indexing [article]

Mihail Stoian and Andreas Kipf and Ryan Marcus and Tim Kraska
2021 arXiv   pre-print
Similar to RadixSpline, PLEX consists of a spline and a (multi-level) radix layer.  ...  However, current learned indexes tend to have many hyperparameters, often do not provide any error guarantees, and are expensive to build. We introduce Practical Learned Index (PLEX).  ...  INTRODUCTION We introduce Practical Learned Index (PLEX). Compared to existing learned indexes, PLEX only has a single hyperparameter (maximum prediction error) and is hence easy to use.  ... 
arXiv:2108.05117v2 fatcat:jde3h4irbvhbfpra7qrzrhxqmm

Shift-Table: A Low-latency Learned Index for Range Queries using Model Correction [article]

Ali Hadian, Thomas Heinis
2021 arXiv   pre-print
As a consequence, querying a learned index on real-world data takes a substantial number of memory lookups, thereby degrading performance.  ...  In this paper, we adopt a different approach for modeling a data distribution that complements the model fitting approach of learned indexes.  ...  The Shift- Table layer is effective in learning almost all distributions even without using models that require training from data, and takes only a single pass over the data points to build the layer  ... 
arXiv:2101.10457v1 fatcat:5go5pdxl4rhyhdzlc2eoi4e6ye

A Lazy Approach for Efficient Index Learning [article]

Guanli Liu and Lars Kulik and Xingjun Ma and Jianzhong Qi
2021 arXiv   pre-print
The synthetic datasets are created to cover a large range of different distributions. Given a new dataset DT, we select the learned index of a synthetic dataset similar to DT, to index DT.  ...  We show a bound over the indexing error when a pre-trained index is selected. We further show how our techniques can handle data updates and bound the resultant indexing errors.  ...  Techniques that learn indices in a single pass such as RadixSpline [8] can be built faster, but they tend to produce sub-optimal indices of large sizes and lower query efficiency.  ... 
arXiv:2102.08081v2 fatcat:qw3ee3fokjexxjnvzcod6jwcha

Benchmarking Learned Indexes [article]

Ryan Marcus, Andreas Kipf, Alexander van Renen, Mihail Stoian, Sanchit Misra, Alfons Kemper, Thomas Neumann, Tim Kraska
2020 arXiv   pre-print
Using four real-world datasets, we demonstrate that learned index structures can indeed outperform non-learned indexes in read-only in-memory workloads over a dense array.  ...  In this work, we present a unified benchmark that compares well-tuned implementations of three learned index structures against several state-of-the-art "traditional" baselines.  ...  In contrast, while a PGM index could theoretically be built in a single pass, the tested implementation of the PGM index builds the initial layer of the index in a single pass, and builds subsequent layers  ... 
arXiv:2006.12804v2 fatcat:76kciq3s3vh4zbbsofgokusoqq

The Case for Distance-Bounded Spatial Approximations [article]

Eleni Tzirita Zacharatou, Andreas Kipf, Ibrahim Sabek, Varun Pandey, Harish Doraiswamy, Volker Markl
2021 arXiv   pre-print
However, approximations are typically only used in a first filtering step to determine a set of candidate spatial objects that may fulfill the query condition.  ...  Furthermore, our approximate techniques employ a distance-based error bound, i.e., a bound on the maximum spatial distance between false (or missing) and exact results which is crucial for meaningful analyses  ...  We employ RadixSpline (RS) as a learned index [15] .  ... 
arXiv:2010.12548v2 fatcat:4snmo6crafho7gujv2vrsdsqeu