pith. machine review for the scientific record. sign in

arxiv: 2604.23451 · v2 · submitted 2026-04-25 · 🪐 quant-ph

Recognition: unknown

Quantum Causal Discovery via Amplitude Estimation of Kullback-Leibler Divergence

Shabnam Sodagari
Authors on Pith no claims yet

Pith reviewed 2026-05-08 08:04 UTC · model grok-4.3

classification 🪐 quant-ph
keywords quantum algorithmscausal discoveryKullback-Leibler divergenceamplitude estimationconditional independence testingPC algorithm
0
0 comments X

The pith

A quantum algorithm estimates clipped Kullback-Leibler divergence with O((L/τ) log(1/δ)) queries, quadratically fewer than classical sampling.

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

The paper introduces QKLA, which encodes a clipped log-density ratio as a quantum amplitude and applies amplitude estimation to recover the corresponding KL expectation. This yields the stated query bound and, when inserted into the PC causal discovery algorithm under per-stratum conditional oracles and a margin condition on decisions, produces an overall Õ(1/(Lτ)) reduction in total oracle queries. The authors prove the scaling, confirm O(1/M) error decay in gate-level simulation, and demonstrate 2.7–7.4× query savings at fixed accuracy on two small networks.

Core claim

QKLA achieves a quadratic precision improvement for estimating clipped KL divergence, needing only O((L/τ)log(1/δ)) queries given coherent oracles and a reversible log-ratio arithmetic oracle; embedding the estimator in the PC algorithm under per-stratum conditional-oracle access and a margin assumption for CI decisions compounds to an Ω̃(1/(Lτ)) reduction in total oracle queries.

What carries the argument

QKLA encodes the clipped log-density ratio as a bounded amplitude and applies amplitude estimation to recover the clipped KL expectation.

If this is right

  • Each conditional independence test requires only O((L/τ)log(1/δ)) queries instead of Θ(1/τ²) samples.
  • Total oracle cost for recovering the causal skeleton scales as Ω̃(1/(Lτ)) times the classical cost.
  • Estimation error decays as O(1/M) where M is the number of queries.
  • On Asia and Synthetic-12 networks the quantum subroutine reaches comparable F1 scores while using 2.7–7.4× fewer queries at the tested precisions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If hardware ever supplies the required coherent oracles, the method could support higher-precision causal graphs in domains where data are expensive or weak dependencies matter.
  • The same amplitude-encoding trick might be reusable for other information measures that appear in constraint-based discovery.
  • Empirical checks of the margin assumption on real data sets would indicate how often the full speedup is realized outside synthetic benchmarks.

Load-bearing premise

Coherent oracle access to the distributions together with a reversible log-ratio arithmetic oracle must exist, and the PC embedding further requires per-stratum conditional oracles plus a margin condition that keeps independence decisions reliable after clipping and estimation error.

What would settle it

A gate-level simulation that plots estimation error versus number of queries M for fixed L and τ and checks whether the slope is exactly 1/M (rather than 1/√M) on a known distribution would settle whether the quadratic scaling holds.

Figures

Figures reproduced from arXiv: 2604.23451 by Shabnam Sodagari.

Figure 1
Figure 1. Figure 1: Experiment 1: state-vector simulation of canonical QAE for view at source ↗
Figure 2
Figure 2. Figure 2: Experiment 2. Classical and quantum query scaling averaged across view at source ↗
Figure 3
Figure 3. Figure 3: Experiment 3. Oracle-model benchmark for PC skeleton-recovery F1 versus total oracle queries on view at source ↗
read the original abstract

Causal discovery from observational data underpins applications in finance, climate modeling, and machine learning. Constraint-based causal discovery reduces structure learning to a sequence of conditional independence (CI) tests, where each test decides independence by estimating conditional mutual information $I(X;Y \mid Z)$ to additive precision $\tau$ and thresholding against it. Classically this requires $\Theta(1/\tau^{2})$ samples per test, a cost that dominates in the high-precision regime typical of weak dependencies. We present QKLA (Quantum Kullback--Leibler Amplitude estimation), a quantum algorithm that encodes a clipped log-density ratio as a bounded amplitude and applies amplitude estimation to recover a clipped KL expectation. Given coherent oracle access to the relevant distributions and a reversible log-ratio arithmetic oracle, QKLA achieves a quadratic precision improvement, needing only $\mathcal{O}((L/\tau)\log(1/\delta))$ queries, where $L$ is the log-ratio clip bound. Under per-stratum conditional-oracle access and a margin assumption for CI decisions, embedding this estimator in the PC algorithm compounds to an $\widetilde{\Omega}(1/(L\tau))$ reduction in total oracle queries. We validate the theory in three experiments. A gate-level state-vector simulation of the full QKLA circuit confirms the predicted $\mathcal{O}(1/M)$ error decay. Across $K=20$ random binary distributions, classical and quantum error scalings match theory to within $0.01$ in slope. In an oracle-model benchmark inside PC on two networks (\textsc{Asia}, 8 nodes; \textsc{Synthetic-12}, 12 nodes), the quantum CI subroutine reaches comparable skeleton-recovery $F_1$ while using $2.7$--$3.2\times$ fewer oracle queries at $\tau = 5\cdot 10^{-3}$ bits and $4.0$--$7.4\times$ fewer at $\tau = 10^{-3}$ bits.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The manuscript introduces QKLA, a quantum algorithm that encodes a clipped log-density ratio as a bounded amplitude and applies amplitude estimation to estimate a clipped version of the conditional KL divergence (hence mutual information) to additive precision τ. It claims a query complexity of O((L/τ) log(1/δ)) under coherent oracles, yielding quadratic improvement over classical Θ(1/τ²) sampling, and shows that embedding the estimator in the PC algorithm produces an overall Õ(1/(Lτ)) reduction in total queries under per-stratum conditional access and a margin assumption on CI decisions. The claims are supported by gate-level simulations confirming 1/M error scaling and benchmark experiments on Asia and Synthetic-12 networks reporting 2.7–7.4× query savings at τ = 10^{-3}–5×10^{-3} while preserving F1 scores.

Significance. If the clipping bias can be controlled so that the effective speedup remains quadratic (or near-quadratic) for the unclipped mutual information required by the PC algorithm, the work would supply a concrete, oracle-model quantum advantage for high-precision constraint-based causal discovery. The gate-level simulations and matching classical/quantum error slopes provide reproducible evidence for the core estimator; the PC benchmarks further demonstrate practical query reduction under the stated assumptions.

major comments (3)
  1. [Abstract] Abstract and the query-complexity statement: the bound O((L/τ) log(1/δ)) is presented with L treated as an independent constant, yet no derivation or lemma shows that the bias |true KL − clipped KL| remains o(τ) for this fixed L. In the worst case over conditional strata whose minimum probability is Θ(τ), the tail contribution where |log(p/q)| > L requires L = Ω(log(1/τ)) to keep bias < τ/2; substituting this scaling removes the quadratic character of the improvement.
  2. [PC embedding discussion] PC-embedding paragraph and the total-reduction claim: the Õ(1/(Lτ)) aggregate saving is derived under an explicit margin assumption that keeps CI decisions reliable after clipping and estimation error. The manuscript does not quantify the fraction of tests in the Asia or Synthetic-12 benchmarks that satisfy this margin, nor does it provide a fallback analysis when the assumption fails.
  3. [Oracle model and assumptions] Oracle-model section: the algorithm requires both coherent access to the joint and conditional distributions and a reversible log-ratio arithmetic oracle. The cost of constructing these oracles from standard quantum oracles for probability distributions is not bounded, so the end-to-end query complexity relative to a classical sampler is not fully established.
minor comments (2)
  1. [Preliminaries] The definition of the clip bound L and the precise form of the reversible log-ratio oracle should be stated explicitly in the main text before the complexity theorem, rather than deferred to an appendix.
  2. [Experiments] Figure captions for the simulation plots should include the exact circuit depth and qubit count used in the state-vector experiments.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the positive assessment and constructive major comments. We address each point in detail below, agreeing where the analysis needs clarification and proposing revisions to the manuscript accordingly.

read point-by-point responses

  1. Referee: [Abstract] Abstract and the query-complexity statement: the bound O((L/τ) log(1/δ)) is presented with L treated as an independent constant, yet no derivation or lemma shows that the bias |true KL − clipped KL| remains o(τ) for this fixed L. In the worst case over conditional strata whose minimum probability is Θ(τ), the tail contribution where |log(p/q)| > L requires L = Ω(log(1/τ)) to keep bias < τ/2; substituting this scaling removes the quadratic character of the improvement.

    Authors: We thank the referee for highlighting this important subtlety. The manuscript presents L as a fixed clip bound chosen for the given distributions, under which the stated quadratic query improvement holds. We acknowledge that, in the worst case where minimum probabilities scale as Θ(τ), maintaining clipping bias o(τ) may require L = Ω(log(1/τ)). We will add a clarifying lemma and discussion in the revision specifying the conditions (e.g., probability mass bounded below by a positive constant independent of τ) under which L remains O(1), while noting the possible logarithmic overhead in the general case. This preserves the core contribution under the paper's assumptions without misrepresenting the worst-case scaling. revision: partial

  2. Referee: [PC embedding discussion] PC-embedding paragraph and the total-reduction claim: the Õ(1/(Lτ)) aggregate saving is derived under an explicit margin assumption that keeps CI decisions reliable after clipping and estimation error. The manuscript does not quantify the fraction of tests in the Asia or Synthetic-12 benchmarks that satisfy this margin, nor does it provide a fallback analysis when the assumption fails.

    Authors: The referee is correct that the overall query-reduction claim relies on the margin assumption. In the reported benchmarks the chosen τ values ensure reliable decisions for the networks tested, but we did not explicitly count the fraction of strata satisfying the margin. We will revise the PC-embedding section to report this fraction for both Asia and Synthetic-12 (computed from the ground-truth distributions) and add a short fallback discussion on the effect of margin violations, such as potential degradation in F1 score or the need for adaptive τ. revision: yes

  3. Referee: [Oracle model and assumptions] Oracle-model section: the algorithm requires both coherent access to the joint and conditional distributions and a reversible log-ratio arithmetic oracle. The cost of constructing these oracles from standard quantum oracles for probability distributions is not bounded, so the end-to-end query complexity relative to a classical sampler is not fully established.

    Authors: This observation is accurate: the analysis isolates the estimator complexity assuming the coherent joint/conditional oracles and reversible log-ratio oracle are given, which is the standard setting for quantum property estimation. We do not provide an end-to-end construction cost from raw data oracles, as that depends on the specific quantum random-access model and is outside the paper's scope (analogous to classical analyses assuming direct sample access). We will add a brief remark clarifying this modeling choice and that the quadratic advantage is with respect to the stated oracle model. revision: no

Circularity Check

0 steps flagged

No circularity: standard amplitude estimation applied to granted oracles

full rationale

The paper derives the O((L/τ)log(1/δ)) query bound for QKLA directly from the standard quantum amplitude estimation theorem once coherent oracles for the distributions and reversible log-ratio arithmetic are assumed as inputs. The PC-algorithm embedding compounds this under an explicit margin assumption for CI decisions. No equation reduces the claimed output to a fitted parameter or self-referential definition, no load-bearing self-citation chain appears, and the clipped-KL construction is presented as an explicit modeling choice rather than a hidden tautology. The derivation chain is therefore self-contained against external quantum query complexity results.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

The central claim rests on standard quantum-oracle assumptions required for amplitude estimation; no free parameters are fitted and no new entities are postulated.

axioms (4)
  • domain assumption Coherent oracle access to the relevant distributions
    Required to prepare the quantum states whose amplitudes encode the clipped log-density ratio.
  • domain assumption Reversible log-ratio arithmetic oracle
    Needed to implement the bounded-amplitude encoding of the log-density ratio inside the quantum circuit.
  • domain assumption Per-stratum conditional-oracle access
    Assumed when embedding the estimator inside the PC algorithm to obtain the overall query reduction.
  • domain assumption Margin assumption for CI decisions
    Ensures that the estimation error after clipping does not flip independence decisions at the chosen threshold τ.

pith-pipeline@v0.9.0 · 5663 in / 1691 out tokens · 89436 ms · 2026-05-08T08:04:42.805433+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

34 extracted references · 20 canonical work pages

  1. [1]

    Spirtes, C

    P. Spirtes, C. Glymour, and R. Scheines,Causation, Prediction, and Search, 2nd ed. Cambridge, MA: MIT Press, 2000

    2000

  2. [2]

    Pearl,Causality: Models, Reasoning, and Inference, 2nd ed

    J. Pearl,Causality: Models, Reasoning, and Inference, 2nd ed. Cambridge University Press, 2009

    2009

  3. [3]

    An algorithm for fast recovery of sparse causal graphs,

    P. Spirtes and C. Glymour, “An algorithm for fast recovery of sparse causal graphs,”Social Science Computer Review, vol. 9, no. 1, pp. 62–72, 1991

    1991

  4. [4]

    Estimation of entropy and mutual information,

    L. Paninski, “Estimation of entropy and mutual information,”Neural Computation, vol. 15, no. 6, pp. 1191–1253, 2003. [Online]. Available: https: //doi.org/10.1162/089976603321780272

    work page doi:10.1162/089976603321780272 2003

  5. [5]

    The adaptive complexity of maximizing a submodular function , booktitle =

    C. L. Canonne, I. Diakonikolas, D. M. Kane, and A. Stewart, “Testing conditional independence of discrete distributions,” inProceedings of the 50th Annual ACM SIGACT Symposium on Theory of Computing. ACM, 2018, pp. 735–748. [Online]. Available: https://doi.org/10.1145/3188745.3188756

    work page doi:10.1145/3188745.3188756 2018

  6. [6]

    Quantum amplitude amplification and estimation.Contemporary Mathematics, 305:53–74, 2002

    G. Brassard, P. Hoyer, M. Mosca, and A. Tapp, “Quantum amplitude amplification and estimation,” 2000. [Online]. Available: https://arxiv.org/abs/quant-ph/0005055

    work page arXiv 2000

  7. [7]

    Quantum speedup of Monte Carlo methods

    A. Montanaro, “Quantum speedup of monte carlo methods,”Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 471, no. 2181, p. 20150301, 2015. [Online]. Available: https://doi.org/10.1098/rspa.2015.0301

    work page doi:10.1098/rspa.2015.0301 2015

  8. [8]

    Local computations with probabilities on graphical structures and their application to expert systems,

    S. L. Lauritzen and D. J. Spiegelhalter, “Local computations with probabilities on graphical structures and their application to expert systems,”Journal of the Royal Statistical Society: Series B (Methodological), vol. 50, no. 2, pp. 157–194, 1988. [Online]. Available: https://doi.org/10.1111/j.2517-6161.1988.tb01721.x

    work page doi:10.1111/j.2517-6161.1988.tb01721.x 1988

  9. [9]

    Quantum query complexity of entropy estimation

    T. Li and X. Wu, “Quantum query complexity of entropy estimation,”IEEE Transactions on Information Theory, vol. 65, no. 5, pp. 2899–2921, 2019. [Online]. Available: https://doi.org/10.1109/TIT.2018.2883306

    work page doi:10.1109/tit.2018.2883306 2019

  10. [10]

    Harrow, and Avinatan Hassidim

    S. Bravyi, A. W. Harrow, and A. Hassidim, “Quantum algorithms for testing properties of distributions,” IEEE Transactions on Information Theory, vol. 57, no. 6, pp. 3971–3981, 2011. [Online]. Available: https://doi.org/10.1109/TIT.2011.2134250

    work page doi:10.1109/tit.2011.2134250 2011

  11. [11]

    T. M. Cover and J. A. Thomas,Elements of Information Theory, 2nd ed. Wiley-Interscience, 2006

    2006

  12. [12]

    Creating superpositions that correspond to efficiently integrable probability distributions

    L. Grover and T. Rudolph, “Creating superpositions that correspond to efficiently integrable probability distributions,” 2002. [Online]. Available: https://arxiv.org/ abs/quant-ph/0208112

    work page Pith review arXiv 2002

  13. [13]

    Distributional property testing in a quantum world , booktitle =

    A. Gily ´en and T. Li, “Distributional property testing in a quantum world,” in11th Innovations in Theoretical Computer Science Conference (ITCS 2020), ser. Leibniz International Proceedings in Informatics (LIPIcs), T. Vidick, Ed., vol. 151. Dagstuhl, Germany: Schloss Dagstuhl – Leibniz-Zentrum f ¨ur Informatik, 2020, pp. 25:1–25:19. [Online]. Available: ...

    work page doi:10.4230/lipics.itcs.2020.25 2020

  14. [14]

    Berry, Andrew M

    D. W. Berry, A. M. Childs, and R. Kothari, “Hamiltonian simulation with nearly optimal dependence on all parameters,” inProceedings of the 56th Annual IEEE Symposium on Foundations of Computer Science (FOCS), 2015, pp. 792–809. [Online]. Available: https://doi.org/10.1109/FOCS.2015.54

    work page doi:10.1109/focs.2015.54 2015

  15. [15]

    Grinko, J

    D. Grinko, J. Gacon, C. Zoufal, and S. Woerner, “Iterative quantum amplitude estimation,”npj Quantum Information, vol. 7, p. 52, 2021. [Online]. Available: https://doi.org/10.1038/s41534-021-00379-1

    work page doi:10.1038/s41534-021-00379-1 2021

  16. [16]

    Suzuki, S

    Y . Suzuki, S. Uno, R. Raymond, T. Tanaka, T. Onodera, and N. Yamamoto, “Amplitude estimation without phase estimation,”Quantum Information Processing, vol. 19, no. 2, p. 75, 2020. [Online]. Available: https://doi.org/10.1007/s11128-019-2565-2

    work page doi:10.1007/s11128-019-2565-2 2020

  17. [17]

    Mitzenmacher and E

    M. Mitzenmacher and E. Upfal,Probability and Com- puting: Randomization and Probabilistic Techniques in Algorithms and Data Analysis, 2nd ed. Cambridge, UK: Cambridge University Press, 2017

    2017

  18. [18]

    Conditional independence testing based on a nearest-neighbor estimator of conditional mutual information,

    J. Runge, “Conditional independence testing based on a nearest-neighbor estimator of conditional mutual information,” inProceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, ser. Proceedings of Machine Learning Research, A. Storkey and F. Perez-Cruz, Eds., vol. 84. PMLR, 2018, pp. 938–947. [Online]. Available:...

    2018

  19. [19]

    Testing conditional independence on discrete data using stochastic complexity,

    A. Marx and J. Vreeken, “Testing conditional independence on discrete data using stochastic complexity,” inProceedings of the Twenty-Second International Conference on Artificial Intelligence and Statistics, ser. Proceedings of Machine Learning Research, K. Chaudhuri and M. Sugiyama, Eds., vol. 89. PMLR, 2019, pp. 496–505. [Online]. Available: https://pro...

    2019

  20. [20]

    Estimating Mutual Information

    A. Kraskov, H. St¨ogbauer, and P. Grassberger, “Estimating mutual information,”Physical Review E, vol. 69, no. 6, p. 066138, 2004. [Online]. Available: https: //doi.org/10.1103/PhysRevE.69.066138

    work page doi:10.1103/physreve.69.066138 2004

  21. [21]

    Geometry of the faithfulness assumption in causal inference,

    C. Uhler, G. Raskutti, P. B¨uhlmann, and B. Yu, “Geometry of the faithfulness assumption in causal inference,”The Annals of Statistics, vol. 41, no. 2, pp. 436–463, 2013. [Online]. Available: https://doi.org/10.1214/12-AOS1080

    work page doi:10.1214/12-aos1080 2013

  22. [22]

    Kernel-based conditional independence test and application in causal discovery,

    K. Zhang, J. Peters, D. Janzing, and B. Sch ¨olkopf, “Kernel-based conditional independence test and application in causal discovery,” inConference on Uncertainty in Artificial Intelligence. Corvallis, OR: AUAI Press, 2011, pp. 804–813. [Online]. Available: https://mlanthology.org/uai/2011/zhang2011uai-kernel/

    2011

  23. [23]

    In: Proceedings of the 2021 ACM-SIAM Symposium on Discrete Algorithms (SODA), pp

    S. Aaronson and P. Rall, “Quantum approximate counting, simplified,” inProceedings of the 3rd Symposium on Simplicity in Algorithms (SOSA). SIAM, 2020, pp. 24–32. [Online]. Available: https://doi.org/10.1137/1. 9781611976014.5

    work page doi:10.1137/1 2020

  24. [24]

    Shende, and Aaron B

    J. Acharya, I. Issa, N. V . Shende, and A. B. Wagner, “Estimating quantum entropy,”IEEE Journal on Selected Areas in Information Theory, vol. 1, no. 2, pp. 454–468, 2020. [Online]. Available: https: //doi.org/10.1109/JSAIT.2020.3015235

    work page doi:10.1109/jsait.2020.3015235 2020

  25. [25]

    Sublinear quantum algorithms for estimating von neumann entropy,

    T. Gur, M.-H. Hsieh, and S. Subramanian, “Sublinear quantum algorithms for estimating von neumann entropy,”

  26. [26]

    Available: https://arxiv.org/abs/2111

    [Online]. Available: https://arxiv.org/abs/2111. 11139

  27. [27]

    A quantum causal discovery algorithm,

    C. Giarmatzi and F. Costa, “A quantum causal discovery algorithm,”npj Quantum Information, vol. 4, p. 17, 2018. [Online]. Available: https://doi.org/10.1038/ s41534-018-0062-6

    2018

  28. [28]

    Estimation of mutual information via quantum kernel method,

    Y . Maeda, H. Kawaguchi, and H. Tezuka, “Estimation of mutual information via quantum kernel method,” 2023. [Online]. Available: https://arxiv.org/abs/2310.12396

    work page arXiv 2023

  29. [29]

    Quantum-enhanced causal discovery for a small number of samples,

    Y . Terada, K. Arai, Y . Tanaka, Y . Maeda, H. Ueno, and H. Tezuka, “Quantum-enhanced causal discovery for a small number of samples,” 2025, later published in Quantum Machine Intelligence 8, Article 36 (2026). [Online]. Available: https://arxiv.org/abs/2501.05007

    work page arXiv 2025

  30. [30]

    Optimal kernel choice for score function-based causal discovery,

    W. Wang, B. Huang, F. Liu, X. You, T. Liu, K. Zhang, and M. Gong, “Optimal kernel choice for score function-based causal discovery,” inProceedings of the 41st International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, R. Salakhutdinov, Z. Kolter, K. Heller, A. Weller, N. Oliver, J. Scarlett, and F. Berkenkamp, Eds., vol. ...

    2024

  31. [31]

    Transformation of quantum states using uniformly controlled rotations,

    M. M ¨ott¨onen, J. J. Vartiainen, V . Bergholm, and M. M. Salomaa, “Transformation of quantum states using uniformly controlled rotations,”Quantum Information & Computation, vol. 5, no. 6, pp. 467–473, 2005. [Online]. Available: https://doi.org/10.26421/QIC5.6-5

    work page doi:10.26421/qic5.6-5 2005

  32. [32]

    Quantum circuits for sparse isometries,

    E. Malvetti, R. Iten, and R. Colbeck, “Quantum circuits for sparse isometries,”Quantum, vol. 5, p. 412, 2021. [Online]. Available: https://doi.org/10.22331/ q-2021-03-15-412

    2021

  33. [33]

    Quantum state preparation with optimal circuit depth: Implementa- tions and applications.Phys

    X.-M. Zhang, T. Li, and X. Yuan, “Quantum state preparation with optimal circuit depth: Implementations and applications,”Physical Review Letters, vol. 129, no. 23, p. 230504, 2022. [Online]. Available: https: //doi.org/10.1103/PhysRevLett.129.230504

    work page doi:10.1103/physrevlett.129.230504 2022

  34. [34]

    Quantum chebyshev’s inequality and applications,

    Y . Hamoudi and F. Magniez, “Quantum chebyshev’s inequality and applications,” in46th International Colloquium on Automata, Languages, and Programming (ICALP 2019), ser. Leibniz International Proceedings in Informatics (LIPIcs), C. Baier, I. Chatzigiannakis, P. Flocchini, and S. Leonardi, Eds., vol. 132. Dagstuhl, Germany: Schloss Dagstuhl – Leibniz- Zent...

    2019