Pith. sign in

REVIEW 5 major objections 5 minor 2 cited by

HERO: Hint-Based Efficient and Reliable Query Optimizer

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Pith's one-line read A hint-based query optimizer that uses no neural network reaches 98.4% of the best possible JOB speedup with zero timeouts and zero degradations, and runs more than five times faster at inference than learned NN-based optimizers.

desk verdict HERO is a solid systems paper with a real new knob (dop) and an honest dynamic-workload evaluation, but the headline non-degradation guarantee only holds in the static lookup setting, not for unseen queries. read the letter →

arxiv 2412.02372 v2 pith:MABOV7LN submitted 2024-12-03 cs.DB cs.AIcs.LG

classification cs.DBcs.AIcs.LG
keywords queryoptimizationhint-basedlearnedoptimizersexecutionplanreusecollisionsdegreeofparallelismreliabilitylocalsearch
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper tries to establish that reliable hint-based query optimization does not require a neural network, and that the usual tension between acceleration and safety can be broken. HERO recommends hints by reusing stored execution plans from queries that share the same default logical plan and similar plan statistics, and it refuses to recommend a hint when the current query is too far from anything previously seen. On the JOB benchmark with static workloads and unlimited training data, HERO achieves 98.4% of the best achievable speedup with zero timeouts and zero degradations, while the neural-network-based baselines degrade or time out even under those ideal conditions. Adding control over the degree of parallelism to the hint space raises the achievable speedup on JOB from roughly 2x to 3x, and HERO realizes that gain. The same components—the context-aware ensemble, the distance metric with a reliability threshold, and a parameterized local search—can also be grafted onto neural-network optimizers to make them safer.

What carries the argument

The central object is the ensemble of context-aware models organized as a graph storage. Vertices are observed execution plans; edges record the hint that transforms one plan into another along with the resulting latency boost. For a new query, HERO generates the default plan, finds the closest model by the distance metric (logical equivalence plus statistics-vector similarity), then follows graph edges to locate promising hints without repeated planning, and only executes the chosen hint combination if its distance to known plans stays below the reliability threshold. The second piece is the parameterized local search procedure: its parameters control which operator families (joins, scans) and which degrees of parallelism are adjusted jointly, how many iterations are run, and whether the Index-Nested-Loop join is handled specially, turning exponential search into a budget-controlled exploration that outperforms greedy search.

What would settle it

Construct two queries with identical default logical plans and statistically similar plan vectors, then apply the hint HERO recommends for the first to the second; if the second query slows down dramatically while the first speeds up, the similarity assumption is broken and the non-degradation guarantee fails for that pair.

Watch

Extended reading notes

Core claim

The paper's central claim is that query latency under a hint depends on the hint only through the execution plan it produces, and that two queries sharing the same default logical plan and similar plan-statistics vectors will respond to the same hint in the same way. On that basis, HERO replaces the neural network with an ensemble of context-aware models: each model is attached to one default plan, stores the custom plans, hints, and observed speedups for queries with that default, and the inference schema reuses previously stored plans from a graph instead of calling the planner repeatedly. A distance metric—infinitely large when logical structures differ, otherwise the relative difference in the plans' statistics vectors—picks the most relevant model and also refuses prediction when the candidate plan is too far from anything seen before, which is what formally provides the non-degradation guarantee. The paper further claims that extending the hint space with degree-of-parallelism control is necessary to unlock the full potential of operator hints, and that a semantics-informed parameterized local search, unlike greedy search, finds near-optimal hint sets within a given exploration budget.

Load-bearing premise

The load-bearing premise is that queries sharing the same default logical plan and similar plan-statistics vectors will respond to the same hints in the same way, so that the distance metric and its reliability threshold truly guarantee non-degradation.

Editorial extensions

If this is right

  • If the reliability guarantee holds, deployment becomes safer: when no close match is found, HERO falls back to the default plan, so an unfamiliar query cannot be slowed down by a bad hint.
  • Because inference reuses stored plans instead of planning under many hint combinations, per-query inference cost drops by more than five times compared to learned NN-based hint optimizers, making hinting practical on large search spaces.
  • Adding degree-of-parallelism control roughly doubles the potential speedup of hint-based optimization on JOB, from about 2x to 3x, and HERO attains that 3x gain.
  • NN-based hint optimizers can inherit part of HERO's safety: using its pruned local search as their inference strategy reduces their inference overhead and degradation rate, although they remain less reliable than HERO on dynamic workloads.
  • The parameterization of local search gives users an explicit dial to trade exploration time against achieved speedup, a flexibility that fixed greedy or exhaustive search does not offer.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The plan-collision principle—identical logical plans can hide opposite responses to the same hint—suggests that any learned optimizer that conditions only on a plan, without query context, faces an inherent accuracy ceiling; clustering by default-plan plus statistics is one way out, and clustering by richer context could improve dynamic-workload behavior.
  • The abstention mechanism could serve as a general safety layer for learned query optimizers: attach the same distance threshold to any hint recommender so that it says 'no hint' when uncertain, independent of whether the recommender is neural or not.
  • Because the experimental measurements were collected on a single database system with a fixed default degree of parallelism, the 3x dop gain and the reliability threshold should be re-tested on another relational engine before the results are taken as engine-agnostic.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. This paper presents HERO, a hint-based query optimizer that replaces the neural-network predictor used in systems such as Bao and AutoSteer with an ensemble of context-aware models organized as a graph, adds a parameterized local-search exploration procedure, and extends the hint space with degree-of-parallelism (dop) control. The central claims are that HERO ensures non-degradation of query latency, trains and infers faster than NN-based alternatives, and reaches near-optimal boosts on static benchmarks: 98.4% of the optimal boost on JOB with 0% timeouts and 0% degradations. The evaluation uses exhaustive plan-latency datasets collected on openGauss for JOB, SQ, and TPC-H, and compares HERO with reimplemented NN-based variants under static, online, and dynamic workload conditions.

Significance. The paper has real strengths: it releases an open dataset and benchmark platform, gives a detailed empirical account of why NN regressors fail on plan-collision-heavy data, and shows that a semantics-informed parameterized local search can find near-optimal hint sets faster than exhaustive search. A working non-neural, interpretable hint optimizer with low degradation risk and dop control would be a useful production-oriented contribution. That said, the current significance is conditional: the headline reliability guarantee is not supported by the dynamic experiments, and the static comparison is close to a lookup benchmark because the training data contains the answer for every test query. These issues materially change what the paper establishes.

major comments (5)
  1. [7.2, Table 8] The static-workload result is a lookup rather than a generalization test. According to Section 7.2, the offline scenario assumes 'all data was observable at training and the exploration budget was unlimited,' and Section 1 describes a dataset containing execution plans and latencies for every query and every hint combination. Under these conditions HERO can reconstruct the best stored hint for each JOB query at inference, so the 98.4% E2E Boost (% of Opt) and 0.0% degradations in Table 8 are expected consequences of exhaustive enumeration, not evidence that HERO predicts reliably for unseen queries. Because the abstract and introduction cite the static results as the basis for the reliability claim, this distinction is load-bearing and must be stated clearly.
  2. [3.3, 4.2, Table 9] The non-degradation guarantee is not established and is contradicted by the dynamic workload results. The constraint in Section 3.3 requires t_i^M + t_i(theta) <= t_i(default) for every query in the support, yet Table 9 shows HERO Local with 4.2% timeouts and 4.2% degradations on the Structure Split. The mechanism intended to enforce this, the distance threshold of Section 4.2, is described only verbally; there is no theorem or calibration experiment showing that the plan-statistics distance bounds the latency difference between a stored plan and a new query's plan. Table 3 is a concrete counterexample at the plan level: queries 6b and 6d share the same default plan, and the same hint theta_ops=126 produces a x1.62 speedup for 6b and a timeout for 6d. Unless the distance metric provably separates such pairs, the threshold cannot deliver the advertised guarantee. The paper's own suggested mitigation, disabling underperforming context models in Section 7.3, is post-hoc and does not satisfy the formal constraint.
  3. [1, Tables 8 and 9] The central latency measurements are reported as single point estimates without repeated runs or confidence intervals. The dataset construction in Section 1 records one latency per query-hint combination on a pre-warmed, exclusively used server, and Tables 8 and 9 contain only single numbers for E2E boost, timeouts, degradations, and inference time. Query latencies on a real DBMS are noisy, so differences such as 64.6% versus 56.2% E2E boost, or inference times of 26.7s versus 58.5s, cannot be assessed for significance. This is especially important because the 'more than five times faster' inference claim and the 3x dop improvement are based on these point estimates.
  4. [7.1, Tables 5-7] The local-search parameterization is evaluated with parameters selected on the same benchmark used to report the result. The paper states that it tested 'tens of thousands of configurations' and then presents the best parameter sets per benchmark in Tables 5-7; these are then used for the comparison in Table 8. Without a separate tuning/validation split or a sensitivity analysis, the comparison between HERO's parameterized local search and the fixed greedy baseline may overstate the advantage, since the greedy baseline does not receive equivalent per-benchmark tuning. The note that the parameters are not universally optimal in Section 7.1 does not address this selection bias.
  5. [5.2, 7] The comparison to existing NN-based systems is performed with the authors' own reimplementations rather than the original systems. Section 7 says 'we implemented a NN architecture that aligns with those in Bao and AutoSteer' and emulates QO-Advisor by pruning a local search. This is a reasonable controlled setup, but the paper's general statements that Bao suffers from 'long training and inference times' and 'unreliable predictions' in Section 2 are then conclusions about this emulation. The inference-time advantage 'more than five times faster' should be reported as relative to the authors' reimplementation, not to the production systems, and the implementation details should be documented to allow a fair comparison.
minor comments (5)
  1. [3.3] The objective and constraint equations contain typesetting errors and missing delimiters, so the formal problem statement is hard to parse; please rewrite them as numbered, self-contained equations.
  2. [4.2] The 'specialized distance metric' is described verbally but the statistics vector and the norm used for the relative difference are never defined; please give the exact formula.
  3. [4.3] The 'super-fast inference' mode based on query templates is not evaluated separately; please report its effectiveness and any accuracy cost relative to the standard inference path.
  4. [1, Abstract] The '3x latency improvement' claim should state in the abstract that it is relative to the default openGauss execution time on JOB and is achieved by adding dop control on top of operator hints; Figure 2 makes this clearer than the abstract does.
  5. [References and notation] There are minor typographical issues: references [1] and [3] contain 'Acessed' instead of 'Accessed', the expression 'theta 0b0011000 = theta_24' is not formatted as a bitmask, and the em-dash in Section 4.3 is followed by a space before a period.

Circularity Check

1 steps flagged · score 6.0 of 10

Headline static-workload result is a lookup, not a prediction: HERO's 98.4%-of-optimal and 0.0%-degradation on JOB are obtained by retrieving exhaustively precomputed hint latencies for the same queries, making the central quality/reliability claim partially circular; the dynamic results give only partial independent support.

  1. fitted input called prediction [Method of Experimental Evaluation (p.1) and §7.2, Table 8]
    "for every query from these benchmarks and for all possible hint combinations, we saved execution plans and their latencies obtained on openGauss DB [1]. This allowed us to replace actual query execution with a simple table lookup. ... The results of comparing HERO and NN-based solutions in conditions, where all data was observable at training and the exploration budget was unlimited, are presented in Table 8."

    HERO's inference is explicitly storage-based: §4.2 stores 'observed performance gains for queries with the same p0', and §4.3 says 'we store directly obtained prior plans and use them to avoid unnecessary planning calls'. In the §7.2 offline scenario every JOB query and every hint combination has already been executed and saved, so for each test query HERO retrieves the best explored hint rather than predicting one for an unseen query. The reported 98.4% of optimal boost and 0.0% degradations therefore equal the best precomputed value in the table with inference overhead subtracted; they are a property of the exhaustive dataset, not of the contextual-transfer machinery that the paper claims provides reliability.

full rationale

The central derivation chain for HERO's headline reliability claim has one genuinely circular link. The static-workload result, which produces the striking 98.4%-of-optimal / 0.0%-degradation numbers, is measured under conditions where 'all data was observable at training and the exploration budget was unlimited', and the paper's own methodology says all queries and all hint combinations were pre-executed and stored so that 'actual query execution' could be replaced by 'a simple table lookup'. Since HERO's ensemble is literally a graph of stored plans and observed gains, the static evaluation reduces to retrieving the stored optimum for the same queries; calling this 'reliable hint recommendation' makes the quality metric definitionally equal to its input. This is not the whole paper, however: the dynamic-workload experiments in Table 9 provide an independent, genuinely predictive setting, and there HERO is safer than the NN baselines (e.g., 0% degradations on three of four splits), which prevents a score of 8 or 10. The paper also contains a non-circular but serious evidentiary gap: §3.3 promises non-degradation for all queries, §4.2 claims the distance threshold ensures this, yet Table 3 shows two queries with the same default plan where the same hint gives x1.62 speedup for 6b and a timeout for 6d, and Table 9 shows HERO itself degrading on the Structure Split; the paper explicitly concedes HERO 'may occasionally arrive at hint sets that slow down queries'. Thus the reliability guarantee is unsupported, and the static evidence for it is circular, even though the underlying retrieval architecture and dynamic comparison have independent content. No load-bearing self-citation or imported uniqueness theorem is present.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claims rest on benchmark-specific tuned local search parameters, an unspecified reliability threshold, and several domain assumptions about how hints affect plans and how plan similarity transfers across queries. No new physical or conceptual entities are introduced.

free parameters (3)
  • Local search configuration per benchmark = JOB: Beta=10.0, Join/Scan/dop/Joinxdop/INL enabled, Iter=infinity; TPCH: Beta=1.0, Join/Scan/dop/Joinxdop enabled…
    The authors scanned tens of thousands of configurations and selected the best per benchmark (Tables 5-7), so the reported optimization ratios are conditional on benchmark-specific tuning.
  • Reliability distance threshold = not specified
    The distance metric in Section 4.2 avoids predictions beyond a 'certain threshold', but the threshold value is never given, making the non-degradation guarantee untestable.
  • Candidate dop values = not specified beyond default of 64 threads
    The degree-of-parallelism grid is not enumerated; Figure 1 implies reducing from 64 threads, but the full set of candidate values is unspecified.
assumptions (4)
  • domain assumption Query latency depends on hints only via their influence on the plan (O2, Section 3.2).
    Used to skip re-executing identical plans and to justify plan-based prediction; may fail if hints affect execution engine behavior beyond plan choice.
  • domain assumption Queries sharing the same default plan and similar plan statistics respond similarly to the same hint (plan collision similarity, Section 4.1).
    This is the transfer assumption underlying the ensemble; if false, reliable predictions for unseen queries are impossible.
  • domain assumption Search space assumptions A1-A4: joins matter more than scans; join and dop joint search is efficient; few iterations suffice; INL on large relations is the common mistake (Section 4.5).
    These motivate the parameterized local search and are validated only empirically on the three benchmarks, not derived.
  • domain assumption openGauss implements hints by heavily overestimating costs of alternative plans, so costs are excluded from plan encoding (Section 5.1).
    This is a DBMS-specific implementation assumption; the encoding may not transfer to other engines.

how reviews work

0 comments
Cite this review

Pith. "Pith review of HERO: Hint-Based Efficient and Reliable Query Optimizer." pith.science (2026). https://pith.science/paper/MABOV7LN

@misc{pith2026241202372,
  author       = {Pith},
  title        = {Pith review of: HERO: Hint-Based Efficient and Reliable Query Optimizer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MABOV7LN}},
  note         = {Machine review of arXiv:2412.02372}
}
read the original abstract

We propose a novel model for learned query optimization which provides query hints leading to better execution plans. The model addresses the three key challenges in learned hint-based query optimization: reliable hint recommendation (ensuring non-degradation of query latency), efficient hint exploration, and fast inference. We provide an in-depth analysis of existing NN-based approaches to hint-based optimization and experimentally confirm the named challenges for them. Our alternative solution consists of a new inference schema based on an ensemble of context-aware models and a graph storage for reliable hint suggestion and fast inference, and a budget-controlled training procedure with a local search algorithm that solves the issue of exponential search space exploration. In experiments on standard benchmarks, our model demonstrates optimization capability close to the best achievable with coarse-grained hints. Controlling the degree of parallelism (query dop) in addition to operator-related hints enables our model to achieve 3x latency improvement on JOB benchmark which sets a new standard for optimization. Our model is interpretable and easy to debug, which is particularly important for deployment in production.

Figures

Figures reproduced from arXiv: 2412.02372 by the authors.

Figure 1
Figure 1. Examples where 𝑑𝑜𝑝 control is necessary to unlock the potential of operator-related hints (query 25a) and to increase it (query 6b). 3.1 Search Space, Θ Hints are typically used to guide an optimizer in the search for better plans [12], but they can also be used to control operation modes such as the degree of parallelism (query 𝑑𝑜𝑝). For example ( [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Influence of hint types on benchmark latency. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Query execution results with the best and worst hints. As one can see, the degradation problem is highly relevant: [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Due to collisions it is difficult to reason about query acceleration in the plan space without having information about the underlying queries. hint options this makes model inference impractical. The second challenge is the collection of a consistent and representativ…
Figure 5
Figure 5. Figure 5: By replacing NN with an ensemble of context-aware models, HERO accelerates inference by reusing plan information [PITH_FULL_IMAGE:figures/full_fig_p005_5.png]
Figure 6
Figure 6. Figure 6: Typical inference schema in existing NN-based solutions for hint-based query optimization. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Learning dynamic on the training data under differ [PITH_FULL_IMAGE:figures/full_fig_p007_7.png]
Figure 8
Figure 8. Figure 8: Using noise instead of plan node information sim [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 9
Figure 9. Figure 9: Four key challenges in training a reliable regressor. Overestimations can occur due to: a) data inconsistency caused [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: The model successfully memorizes the training [PITH_FULL_IMAGE:figures/full_fig_p009_10.png]
Figure 12
Figure 12. Figure 12: Poor generalization on OOD data is, in part, due [PITH_FULL_IMAGE:figures/full_fig_p009_12.png]
Figure 13
Figure 13. Figure 13: Results of online scenario simulation on JOB [PITH_FULL_IMAGE:figures/full_fig_p011_13.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Learned Offline Query Planning via Bayesian Optimization

    cs.DB 2025-02 conditional novelty 6.0 of 10

    BayesQO combines variational autoencoders and Bayesian optimization with censored timeouts to discover faster join-order plans offline for repetitive analytic workloads.

  2. Training-Free Query Optimization via LLM-Based Plan Similarity

    cs.DB 2025-06 conditional novelty 5.0 of 10

    LLM-PM, a training-free system that transfers optimizer hints from nearest-neighbour LLM plan embeddings, reports a 21.1% total runtime reduction on JOB-CEB over openGauss.

Reference graph

Works this paper leans on

27 extracted references · 19 canonical work pages · cited by 2 Pith papers

  1. [1]

    openGauss

    [n.d.]. openGauss. https://opengauss.org/en/aboutUs/. Acessed: 2024-10-10

  2. [2]

    pg_hint_plan Documentation

    [n.d.]. pg_hint_plan Documentation. https://pg-hint-plan.readthedocs.io/en/ latest/. Accessed: 2024-10-10

  3. [3]

    PostgreSQL

    [n.d.]. PostgreSQL. https://www.postgresql.org/about/. Acessed: 2024-10-10

  4. [4]

    Christoph Anneser, Nesime Tatbul, David Cohen, Zhenggang Xu, Prithviraj Pandian, Nikolay Laptev, and Ryan Marcus. 2023. Autosteer: Learned query 12 optimization for any sql database. Proceedings of the VLDB Endowment 16, 12 (2023), 3515–3527

  5. [5]

    Michael M Bronstein, Joan Bruna, Yann LeCun, Arthur Szlam, and Pierre Van- dergheynst. 2017. Geometric deep learning: going beyond euclidean data. IEEE Signal Processing Magazine 34, 4 (2017), 18–42

  6. [6]

    Surajit Chaudhuri. 1998. An overview of query optimization in relational systems. In Proceedings of the seventeenth ACM SIGACT-SIGMOD-SIGART symposium on Principles of database systems . 34–43

  7. [7]

    Surajit Chaudhuri, Umeshwar Dayal, and Vivek Narasayya. 2011. An overview of business intelligence technology. Commun. ACM 54, 8 (2011), 88–98

  8. [8]

    Transaction Processing Council. [n.d.]. TPC-H. https://www.tpc.org/tpch/. Accessed: 2024-10-10

Show all 27 references
  1. [9]

    Andreas Kipf, Thomas Kipf, Bernhard Radke, Viktor Leis, Peter Boncz, and Alfons Kemper. 2018. Learned cardinalities: Estimating correlated joins with deep learning. arXiv preprint arXiv:1809.00677 (2018)

  2. [10]

    Viktor Leis, Andrey Gubichev, Atanas Mirchev, Peter Boncz, Alfons Kemper, and Thomas Neumann. 2015. How good are query optimizers, really? Proceedings of the VLDB Endowment 9, 3 (2015), 204–215

  3. [11]

    Guy Lohman. 2014. Is query optimization a “solved” problem. In Proc. Workshop on Database Query Optimization , Vol. 13. Oregon Graduate Center Comp. Sci. Tech. Rep, 10

  4. [12]

    Ryan Marcus, Parimarjan Negi, Hongzi Mao, Nesime Tatbul, Mohammad Al- izadeh, and Tim Kraska. 2020. Bao: Learning to steer query optimizers. arXiv preprint arXiv:2004.03814 (2020)

  5. [13]

    Ryan Marcus, Parimarjan Negi, Hongzi Mao, Nesime Tatbul, Mohammad Al- izadeh, and Tim Kraska. 2021. Bao: Making learned query optimization practical. In Proceedings of the 2021 International Conference on Management of Data . 1275– 1288

  6. [14]

    Ryan Marcus, Parimarjan Negi, Hongzi Mao, Chi Zhang, Mohammad Alizadeh, Tim Kraska, Olga Papaemmanouil, and Nesime Tatbul. 2019. Neo: A learned query optimizer. arXiv preprint arXiv:1904.03711 (2019)

  7. [15]

    Ryan Marcus and Olga Papaemmanouil. 2018. Deep reinforcement learning for join order enumeration. In Proceedings of the First International Workshop on Exploiting Artificial Intelligence Techniques for Data Management . 1–4

  8. [16]

    Lili Mou, Ge Li, Lu Zhang, Tao Wang, and Zhi Jin. 2016. Convolutional neural net- works over tree structures for programming language processing. In Proceedings of the AAAI conference on artificial intelligence

  9. [17]

    Parimarjan Negi, Matteo Interlandi, Ryan Marcus, Mohammad Alizadeh, Tim Kraska, Marc Friedman, and Alekh Jindal. 2021. Steering query optimizers: A practical take on big data workloads. In Proceedings of the 2021 International Conference on Management of Data . 2557–2569

  10. [18]

    Parimarjan Negi, Ryan Marcus, Andreas Kipf, Hongzi Mao, Nesime Tatbul, Tim Kraska, and Mohammad Alizadeh. 2021. Flow-loss: Learning cardinality estimates that matter. arXiv preprint arXiv:2101.04964 (2021)

  11. [19]

    Parimarjan Negi, Ryan Marcus, Hongzi Mao, Nesime Tatbul, Tim Kraska, and Mohammad Alizadeh. 2020. Cost-guided cardinality estimation: Focus where it matters. In 2020 IEEE 36th International Conference on Data Engineering Workshops (ICDEW). IEEE, 154–157

  12. [20]

    Jennifer Ortiz, Magdalena Balazinska, Johannes Gehrke, and S Sathiya Keerthi

  13. [21]

    Tarique Siddiqui, Alekh Jindal, Shi Qiao, Hiren Patel, and Wangchao Le. 2020. Cost models for big data query processing: Learning, retrofitting, and our findings. In Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data. 99–113

  14. [22]

    Ji Sun, Jintao Zhang, Zhaoyan Sun, Guoliang Li, and Nan Tang. 2021. Learned cardinality estimation: A design space exploration and a comparative evaluation. Proceedings of the VLDB Endowment 15, 1 (2021), 85–97

  15. [23]

    Lucas Woltmann, Jerome Thiessat, Claudio Hartmann, Dirk Habich, and Wolf- gang Lehner. 2023. Fastgres: Making learned query optimizer hinting effective. Proceedings of the VLDB Endowment 16, 11 (2023), 3310–3322

  16. [24]

    Zongheng Yang, Wei-Lin Chiang, Sifei Luan, Gautam Mittal, Michael Luo, and Ion Stoica. 2022. Balsa: Learning a query optimizer without expert demonstrations. In Proceedings of the 2022 International Conference on Management of Data. 931–944

  17. [25]

    Wangda Zhang, Matteo Interlandi, Paul Mineiro, Shi Qiao, Nasim Ghazanfari, Karlen Lie, Marc Friedman, Rafah Hosn, Hiren Patel, and Alekh Jindal. 2022. Deploying a steered query optimizer in production at microsoft. In Proceedings of the 2022 International Conference on Managem...

  18. [26]

    Yue Zhao, Zhaodonghui Li, and Gao Cong. 2023. A Comparative Study and Com- ponent Analysis of Query Plan Representation Techniques in ML4DB Studies. Proceedings of the VLDB Endowment 17, 4 (2023), 823–835. 13

  19. [2019]

    arXiv preprint arXiv:1905.06425 (2019)

    An empirical analysis of deep learning for cardinality estimation. arXiv preprint arXiv:1905.06425 (2019)

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.