Can Learned Models Replace Hash Functions?

Summary: Evaluates learned models as hash functions across chaining, probing, and cuckoo hashing. They cut probe latency 1.4× and non-partitioned join time 28% on favorable distributions, but can hurt when distributions are unsuitable. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: h97057314b98389f5
Venue: VLDB
Year: 2023
Pagerank: 6.3452789e-05
Overall Rank: 4,915 | 67.06%
DOI: 10.14778/3570690.3570702
PDF: Download (CC BY-NC-ND 4.0)

Incoming Non-self Citations Over Time

Authors

1. Ibrahim Sabek (Massachusetts Institute of Technology)
2. Kapil Vaidya (Massachusetts Institute of Technology)
3. Dominik Horn (Technical University of Munich)
4. Andreas Kipf (Massachusetts Institute of Technology)
5. Michael Mitzenmacher (Harvard University)
6. Tim Kraska (Massachusetts Institute of Technology)

BibTeX Citation

@article{sabek_vldb23,
        title = {{Can Learned Models Replace Hash Functions?}},
        author = {Sabek, Ibrahim and Vaidya, Kapil and Horn, Dominik and Kipf, Andreas and Mitzenmacher, Michael and Kraska, Tim},
        journal = {PVLDB},
        series = {{VLDB} '23},
        volume = {16},
        number = {3},
        pages = {532--545},
        doi = {10.14778/3570690.3570702},
        url = {https://doi.org/10.14778/3570690.3570702},
        year = {2023}
}

Incoming Citations (Sorted by Pagerank)

Showing 8 of 8 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
3,632	Learned Index: A Comprehensive Experimental Evaluation	2023	VLDB	7.1274345e-05
7,439	Is Perfect Hashing Practical for OLAP Systems?	2024	CIDR	5.5122587e-05
8,419	The Case for Learned In-Memory Joins	2023	VLDB	5.3198816e-05
10,552	LINE: A Learned Index with Group-Enhanced Leaves and Cache-Optimized Inner Tree	2026	SIGMOD	4.9605343e-05
10,762	The Case For Language Model Approximated LIKE Predicate	2026	SIGMOD	4.9605343e-05
11,140	Learned Static Function Data Structures	2026	VLDB	4.9605343e-05
11,393	Not Small Enough? SegPQ: A Learned Approach to Compress Product Quantization Codebooks	2025	VLDB	4.9605343e-05
11,574	FairHash: A Fair and Memory/Time-efficient Hashmap	2024	SIGMOD	4.9605343e-05

Outgoing Citations (Sorted by Pagerank)

Showing 18 of 18 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
40	The Case for Learned Index Structures	2018	SIGMOD	0.00046249672
84	Learned Cardinalities: Estimating Correlated Joins with Deep Learning	2019	CIDR	0.00035920657
355	Bao: Making Learned Query Optimization Practical	2021	SIGMOD	0.00020146245
427	ALEX: An Updatable Adaptive Learned Index	2020	SIGMOD	0.00018429918
461	The PGM-index: a fully-dynamic compressed learned index with provable worst-case bounds	2020	VLDB	0.00017835351
834	Benchmarking Learned Indexes	2021	VLDB	0.00013539098
870	Learning Multi-dimensional Indexes	2020	SIGMOD	0.0001332034
906	HOT: A Height Optimized Trie Index for Main-Memory Database Systems	2018	SIGMOD	0.00013129817
1,119	A Comprehensive Study of Main-Memory Partitioning and its Application to Large-Scale Comparison- and Radix-Sort	2014	SIGMOD	0.00011927356
1,190	Tsunami: A Learned Multi-dimensional Index for Correlated Data and Skewed Workloads	2021	VLDB	0.00011564169
1,213	A Seven-Dimensional Analysis of Hashing Methods and its Implications on Query Processing	2016	VLDB	0.00011492443
1,269	An Experimental Comparison of Thirteen Relational Equi-Joins in Main Memory	2016	SIGMOD	0.00011236561
1,445	LISA: A Learned Index Structure for Spatial Data	2020	SIGMOD	0.00010607118
1,884	Effectively Learning Spatial Indices	2020	VLDB	9.422145e-05
2,814	Exploiting Coroutines to Attack the "Killer Nanoseconds"	2018	VLDB	7.9550699e-05
4,773	Cuckoo Index: A Lightweight Secondary Index Structure	2020	VLDB	6.4102054e-05
4,797	LSched: A Workload-Aware Learned Query Scheduler for Analytical Database Systems	2022	SIGMOD	6.3978116e-05
5,642	Revisiting Reuse in Main Memory Database Systems	2017	SIGMOD	6.0370517e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	834	Benchmarking Learned Indexes	2021	VLDB
2	11,931	New Wine in an Old Bottle: Data-Aware Hash Functions for Bloom Filters	2022	VLDB
3	40	The Case for Learned Index Structures	2018	SIGMOD
4	360	Design and Evaluation of Main Memory Hash Join Algorithms for Multi-core CPUs	2011	SIGMOD
5	3,632	Learned Index: A Comprehensive Experimental Evaluation	2023	VLDB
6	8,419	The Case for Learned In-Memory Joins	2023	VLDB
7	7,782	Adaptive Factorization Using Linear-Chained Hash Tables	2025	CIDR
8	3,552	A General and Efficient Querying Method for Learning to Hash	2018	SIGMOD
9	9,034	Entropy-Learned Hashing: Constant Time Hashing with Controllable Uniformity	2022	SIGMOD
10	1,213	A Seven-Dimensional Analysis of Hashing Methods and its Implications on Query Processing	2016	VLDB