Pith. sign in

REVIEW 2 major objections 4 minor 78 references

MATERO-RCA jointly optimizes root sets, root-effect modes, and counterfactual trajectories to resolve industrial alarms, achieving 97.8% mean Set F1 and 94.3% Exact Set across five dataset groups.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

MATERO-RCA jointly optimizes root set, root-effect modes, and repaired counterfactual trajectories, with a MILP-accelerated best-bound search that reports strong benchmark RCA results.

T0 review reviewed 2026-08-03 challenge →

load-bearing objection A genuinely new joint root-set/mode/counterfactual optimization with a certified search, but the empirical headline is conditional on a graph-completeness assumption and missing artifacts. the 2 major comments →

arxiv 2607.29092 v1 pith:DCT3TQBQ submitted 2026-07-31 eess.SP

MATERO-RCA: Mode-Aware Trajectory-Level Energy-Based Root-Set Optimization for Industrial Root Cause Analysis

Chengyu Tao , Chunxi Huang , Runquan Xiao This is my paper
classification eess.SP
keywords root cause analysisindustrial time seriesenergy-based modelcounterfactual trajectoriesbest-bound searchtemporal compatibilitymixed-integer linear program
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

The paper proposes a root cause analysis method for industrial time series where a signal can look normal in isolation but violate its operating context. Instead of ranking anomalies or simulating forward effects, MATERO-RCA jointly selects a set of root variables, assigns each root either an observation-only or physical-propagation mode, and optimizes counterfactual trajectories that would resolve the active alarms. The graph-wide objective combines alarm resolution with temporal compatibility across local causal relations, and an exact MILP search with a residual-cover lower bound certifies that no better root set was missed under the fixed inner solver. On five simulated and real industrial dataset groups, the method reports the highest set-level accuracy across all baselines.

Core claim

The paper introduces MATERO-RCA, which treats root cause identification as minimizing an energy function over counterfactual trajectories rather than ranking anomalies or performing candidate-wise forward simulation. A Temporal Compatibility Network (CompatNet) learns calibrated energies for local causal relations and alarm-parent contexts, and a Counterfactual Repair Network (RepairNet) initializes mode-aware counterfactual trajectories that are refined by gradient descent under the same objective. The outer search minimizes a graph-wide objective with a cardinality penalty, and a residual-cover lower bound enables certified best-bound search over the finite root–mode space. The paper repor

What carries the argument

The central machinery is an energy-based objective J(x) = Φ_A(x) + λ Φ_C(x) with two learned components: CompatNet provides calibrated compatibility energies for local causal relations (Φ_C) and alarm-parent contexts (Φ_A), while RepairNet generates diverse counterfactual trajectory proposals that are refined by straight-through gradient optimization. The outer search uses a residual-cover lower bound, exactly encoded as a binary MILP, to prioritize unseen root sets and certify fixed-oracle outer optimality.

Load-bearing premise

The causal graph provided must be correct for root-to-alarm reachability, because the search restricts candidates to ancestors of top-level active alarms and has no way to propose a true root that lies outside that ancestor set.

What would settle it

Construct an event where the true root is not an ancestor of any top-level active alarm in the given graph (e.g., a latent cause) and show that MATERO-RCA reports a root set that omits the true root while achieving a low objective value; this would directly violate the paper's central claim that the root set is recoverable from the graph-factored objective.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • RCA becomes a trajectory-level optimization problem rather than anomaly-first ranking, directly handling multi-root, multi-alarm events.
  • The certified best-bound search guarantees that no better root set exists within the admissible root–mode space, given the learned energy model and fixed inner solver.
  • The method handles observation-only root effects, which previous methods excluded, by explicitly modeling root-effect modes.
  • The ablations show that alarm resolution and temporal compatibility provide complementary evidence; removing either term degrades set recovery.

Where Pith is reading between the lines

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

  • The load-bearing assumption is that the supplied causal graph correctly captures all root-alarm reachability; the method cannot propose a root that is not an ancestor of a top-level active alarm in the graph.
  • The approach could be extended to incorporate graph uncertainty by jointly optimizing over a distribution of causal graphs or by adding graph-structure refinement.
  • The certified search's guarantees are relative to the learned compatibility model, not the true physical system; failures attributed to out-of-distribution propagation suggest a need for open-set response models.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 4 minor

Summary. MATERO-RCA proposes an RCA method that, given a causal graph G and normal training windows, jointly optimizes a root set, per-root effect modes (observation-only vs physical propagation), and counterfactual trajectories. CompatNet supplies calibrated energies for alarm-context plausibility and local parent–child compatibility; RepairNet proposes diverse counterfactuals refined by gradient descent under the same objective. The outer combinatorial search is driven by a residual-cover lower bound encoded as an exact MILP, with anytime optimality guarantees for a fixed inner solver. Experiments on one real industrial dataset (causRCA), three synthetic generators, and the Quadruple-Tank simulation report mean Set F1 97.8% and Exact Set 94.3%, together with ablations, sensitivity studies, and failure analyses. The appendix contains proofs of Propositions 1–2, Theorem 1, and Corollary 1, with the guarantees explicitly scoped to the finite admissible root–mode space and the fixed inner solver.

Significance. The proposed framework is a coherent and nontrivial integration of learned trajectory-level compatibility, mode-aware counterfactual generation, and certified combinatorial search. If the empirical claims hold, this is a meaningful advance over score- and rollout-based temporal RCA baselines. The paper is commendably careful in hedging its theoretical claims: the proofs are in the appendix, the certificates are stated for a fixed inner solver, and the authors include failure analysis and extensive sensitivity studies rather than only favorable tables. The principal weakness is that the central empirical claim is conditional on the supplied graph being complete for root–alarm reachability, a condition that is not verified on the real dataset; in addition, no code/data or uncertainty quantification accompany the headline numbers.

major comments (2)
  1. [Problem Setting and Notation; S4.4; S7.5] The candidate restriction C = ∪_{A_j∈A^⊤} an(A_j) is load-bearing. Any true root that is not an ancestor of a top-level active alarm under the supplied graph is unrecoverable, regardless of CompatNet/RepairNet quality. The paper itself disclaims causal completeness outside X in S4.4, and the Conclusion attributes remaining failures to causal-graph misspecification. For the synthetic groups the roots are in C by construction; for the only real group, causRCA, the paper does not report whether every annotated root lies in C, nor whether the annotations were produced independently of the expert graph. The S7.5 perturbation study explicitly keeps 'declared root-to-alarm reachability' fixed, so it cannot probe the candidate ceiling. The headline claims should therefore be qualified as conditional on graph completeness, and the authors should add a per-event check that annotated roots are in C
  2. [Experiments: Table 1; Implementation] The central empirical claim rests on single point estimates. Table 1 (and Tables 2–3) report no error bars, confidence intervals, or per-seed variation; the appendix's sensitivity experiments fix model seed 1 (e.g., S7.5, S7.4). With test sets of 25–50 events, a difference of one or two events changes the percentages by 2–4 points, so the statement that MATERO-RCA 'outperforms all baselines' needs uncertainty quantification. No code or data are provided, making the results in Tables 1–2 impossible to reproduce independently. I do not regard this as a correctness error, but it is necessary support for the claimed state-of-the-art performance.
minor comments (4)
  1. [Theorem 1, Eq. (11)] The inequality 0 ≤ bU_t − bU⋆ ≤ [bU_t − B_t]_+ is stated for every iteration t, but at t = 0, bU_0 = +∞ and the expression is not meaningful. The proof correctly notes that the finite anytime gap starts at t = 1; the theorem statement should be rephrased accordingly.
  2. [Table 1 and Table S4] The main text does not explain that the causRCA column is a macro average over four overlapping views (Probe, Coolant, Hydraulics, Full) of the same 100 physical events. This should be stated next to Table 1 to avoid the impression that the column aggregates independent test events.
  3. [Table 2] Some entries use inconsistent abbreviations ('Stable.', 'MATERO.', 'Smooth.'). Use the full method names or a uniform abbreviation scheme.
  4. [General] A data/code availability statement is missing. Given the diversity of datasets and the dependence on many hyperparameters (ε, λ, γ, ρ_C, ρ_A, bin counts, etc.), releasing code and trained configurations would substantially strengthen reproducibility.

Circularity Check

0 steps flagged

No significant circularity: the derivation is self-contained, root labels and effect modes are not used to fit parameters, and the search certificates are proven and explicitly scoped.

full rationale

The paper's central derivation does not reduce to its inputs. CompatNet and RepairNet are trained only on normal windows; the training protocol explicitly states that 'Fault labels, root annotations, and test data or metrics are excluded from training, calibration, and checkpoint selection.' Root labels are used only for evaluation, not for fitting the energy or the repair initializer. The objective J in Eq. (1) is defined from learned calibrated energies (Eqs. 3-5) and is then minimized over counterfactual trajectories for each root-mode assignment; this is a self-supervised scoring criterion, not a fitted proxy for the annotated root set. Residual-cover best-bound search is supported by proofs (Proposition 1, Proposition 2, Theorem 1, Corollary 1) rather than assumed equivalence, and S4.4 explicitly limits the certificate scope: 'All certificates concern the finite admissible root–mode space for the provided graph, frozen checkpoints, and fixed inner-solver settings and proposal-noise draws. They do not certify the global optimum of the continuous inner problem, the correctness of the learned compatibility model, or causal completeness outside X.' The only load-bearing premise is that the provided causal graph is complete for root-to-alarm reachability, so candidates are restricted to C = union of ancestors of top-level active alarms. That is an external-assumption limitation, not a circularity: the paper does not define the true root set in terms of C, and S7.5's perturbation study, while limited, does test some graph misspecification. No self-citation chain is load-bearing; the cited works (HiGHS, GRU, straight-through estimator, quantile calibration, etc.) are standard and independent, and the paper does not invoke a prior uniqueness theorem by the same authors to forbid alternatives. The reader's note that RepairNet is trained with the same J used at inference reflects self-referentiality in an amortized initializer, but the final root-set decision is evaluated against independent root annotations on held-out test events, so the claimed result is not equal to a training input by construction.

Axiom & Free-Parameter Ledger

9 free parameters · 6 axioms · 1 invented entities

The ledger shows that the method adds two trained networks and many hyperparameters, and depends on strong prior knowledge: a correct causal graph, alarm-free normal training data, irreducible roots, and physical propagation that stays within learned normal support. The paper is honest about some of these dependencies, but the high benchmark scores are conditional on all of them.

free parameters (9)
  • cardinality cap K = 3
    Outer search is restricted to root sets of size 1..K; if a true multi-root event has more than K roots, it cannot be recovered.
  • energy balance λ = 0.5
    Balances unresolved alarm energy vs local compatibility energy in J = Φ_A + λΦ_C; no principled selection is given.
  • cardinality penalty γ = 0.25
    Controls the trade-off between inner energy and the number of roots in U; relies on the unproven irreducibility assumption.
  • propagation threshold ε = 0.01
    Defines the descendant scope of p-mode roots; the paper reports that ExactSet is sensitive to ε.
  • calibration quantiles ρ_C / ρ_A = 0.99 / 0.95
    Set zero energy within estimated normal support; appendix S7.3 shows ρ_A sensitivity.
  • continuous bin count (QT) = 7 (best of {5,6,7,8})
    S7.4 reports that seven bins is the only setting with perfect A@1 and C@3; the choice is made after comparing test-set scores.
  • refinement schedule = 50 AdamW steps, lr 0.2, τ annealed 1→0.25
    Gradient refinement settings define the fixed inner solver that all certificates refer to.
  • proposal count and RepairNet losses = L=4, κ=0.05, α=1, β=0.01
    Multi-proposal generation and diversity weighting; no sensitivity is reported for α, β, or κ.
  • baseline selection α = 0.01 (Holm)
    Appendix S7.2 evaluates Holm α values and fixes the one with the highest covered-set ES; not tuned per method, but selected using test metrics.
axioms (6)
  • domain assumption The provided causal graph G is correct and complete for root-alarm reachability; all true roots lie in C = ∪_{A_j ∈ A^⊤} an(A_j).
    Problem Setting: 'Since non-ancestors cannot causally reach a top-level active alarm, candidates are restricted to C=...' If the graph is misspecified or a latent cause exists, true roots are excluded before search; the paper's conclusion attributes remaining failures to causal-graph misspecification.
  • ad hoc to paper The generating root set is irreducible and has cardinality ≤ K; every lower-cardinality alternative raises the inner value more than it reduces the cardinality penalty.
    Methodology after Eq. (2): 'We assume the generating root set is irreducible...' This is needed for the γ|R| penalty to recover the true set; it is not externally testable.
  • domain assumption Normal training windows are alarm-free and contain the same local causal mechanisms; calibrated quantiles of their energies define the normal support.
    CompatNet training and Eq. (5). If normal data omit regimes that occur during faults, p-mode propagation can be outside learned support, as admitted in the failure analysis.
  • domain assumption An active alarm is resolved iff its alarm-parent context returns to the learned normal joint-state support.
    Eqs. (4)-(5) model the alarm-parent context rather than the alarm variable itself; this equivalence is central to the Φ_A term.
  • domain assumption Discrete binning of continuous signals preserves fault-relevant information.
    S1.1 maps continuous signals to fixed bins; S7.4 shows two QT events are missed because clean and faulty values fall in the same bin.
  • domain assumption Physical-propagation root effects follow response patterns supported by normal training data.
    Failure Analysis: seven of nine ExactSet errors are predefined OOD events that violate this assumption, causing the method to add one causal neighbor.
invented entities (1)
  • Root-effect modes o and p (observation-only vs physical-propagation) no independent evidence
    purpose: Split each root's causal effect into a recorded-trajectory-only mode or a downstream-propagating mode so the optimizer can explain contextual anomalies without an explicit SCM.
    The modes are auxiliary and the paper states they 'need not be uniquely identifiable' (Problem Setting). They introduce an extra modeling degree of freedom rather than a measured physical quantity.

reviewed 2026-08-03 · how reviews work

0 comments
Cite this review

Pith. "Pith review of MATERO-RCA: Mode-Aware Trajectory-Level Energy-Based Root-Set Optimization for Industrial Root Cause Analysis." pith.science (2026). https://pith.science/paper/DCT3TQBQ

@misc{pith2026260729092,
  author       = {Pith},
  title        = {Pith review of: MATERO-RCA: Mode-Aware Trajectory-Level Energy-Based Root-Set Optimization for Industrial Root Cause Analysis},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DCT3TQBQ}},
  note         = {Machine review of arXiv:2607.29092}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Root cause analysis (RCA) for contextual anomalies in industrial time series is challenging because responses depend jointly on control commands, operating states, and coupled physical variables. A response can appear marginally normal yet violate its operating context. Events may involve multiple roots and alarms, with each root assigned an observation-only effect confined to its recorded trajectory or a physical-propagation effect on descendants. We propose Mode-Aware Trajectory-Level Energy-Based Root-Set Optimization for Root Cause Analysis (MATERO-RCA), which jointly optimizes a root set, root-effect modes, and auxiliary counterfactual trajectories. Its graph-wide objective combines alarm resolution with temporal compatibility across local causal relations. A Temporal Compatibility Network(CompatNet) maps parent-conditioned trajectory likelihoods to calibrated compatibility energies. A Counterfactual Repair Network (RepairNet) initializes mode-aware counterfactual trajectories for objective-directed gradient refinement. An exact mixed-integer linear program minimizes a residual-cover lower bound, enabling certified best-bound search over the finite admissible root--mode space under the fixed inner solver. Experiments on simulated and real industrial datasets demonstrate superior RCA performance over representative baselines.

Figures

Figures reproduced from arXiv: 2607.29092 by Chengyu Tao, Chunxi Huang, Runquan Xiao.

Figure 1
Figure 1. Figure 1: Illustrative industrial event in which roots with [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: MATERO-RCA integrates graph-factored temporal compatibility and mode-aware counterfactual trajectory optimiza [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: Exhaustive vs. best-bound search. Best-bound [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗

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

78 extracted references · 6 canonical work pages · 5 internal anchors

  1. [1]

    K.; Ez-Zejjari, I.; and Zan, L

    Assaad, C. K.; Ez-Zejjari, I.; and Zan, L. 2023. Root Cause Identification for Collective Anomalies in Time Series Given an Acyclic Summary Causal Graph with Loops. In Proceedings of the 26th International Conference on Artificial Intelligence and Statistics, volume 206 of Proceedings of Machine Learning Research, 8395--8404. PMLR

    2023

  2. [2]

    Z.; and Koltun, V

    Bai, S.; Kolter, J. Z.; and Koltun, V. 2018. An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling. arXiv preprint arXiv:1803.01271

    Pith/arXiv arXiv 2018

  3. [3]

    Bengio, Y.; L \'e onard, N.; and Courville, A. 2013. Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation. arXiv preprint arXiv:1308.3432

    Pith/arXiv arXiv 2013

  4. [4]

    Benjamini, Y.; and Hochberg, Y. 1995. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society: Series B (Methodological), 57(1): 289--300

    1995

  5. [5]

    Budhathoki, K.; Minorics, L.; Bl \"o baum, P.; and Janzing, D. 2022. Causal Structure-Based Root Cause Analysis of Outliers. In Chaudhuri, K.; Jegelka, S.; Song, L.; Szepesvari, C.; Niu, G.; and Sabato, S., eds., Proceedings of the 39th International Conference on Machine Learning, volume 162 of Proceedings of Machine Learning Research, 2357--2369. PMLR

    2022

  6. [6]

    U.; Aubet, F.-X.; Flunkert, V.; and Gasthaus, J

    Carmona, C. U.; Aubet, F.-X.; Flunkert, V.; and Gasthaus, J. 2022. Neural Contextual Anomaly Detection for Time Series. In Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence, 2843--2851

    2022

  7. [7]

    Chandola, V.; Banerjee, A.; and Kumar, V. 2009. Anomaly Detection: A Survey. ACM Computing Surveys, 41(3): 1--58

    2009

  8. [8]

    Cho, K.; van Merri \"e nboer, B.; Gulcehre, C.; Bahdanau, D.; Bougares, F.; Schwenk, H.; and Bengio, Y. 2014. Learning Phrase Representations Using RNN Encoder--Decoder for Statistical Machine Translation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, 1724--1734. Association for Computational Linguistics

    2014

  9. [9]

    Chraim, F.; Zhang, J.; Janzing, D.; Song, X.; Faloutsos, C.; and Evans, J. 2026. NetCause : Counterfactual Learning for Root Cause Analysis in Large-Scale Networks. arXiv:2606.13543

    Pith/arXiv arXiv 2026

  10. [10]

    Cortes, A.; Rehm, R.; and Letzelter, V. 2025. Winner-Takes-All for Multivariate Probabilistic Time Series Forecasting. In Proceedings of the 42nd International Conference on Machine Learning, volume 267 of Proceedings of Machine Learning Research, 11288--11312. PMLR

    2025

  11. [11]

    Deng, A.; and Hooi, B. 2021. Graph Neural Network-Based Anomaly Detection in Multivariate Time Series. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, 4027--4035

    2021

  12. [12]

    Han, X.; Absar, S.; Zhang, L.; and Yuan, S. 2025. Root Cause Analysis of Anomalies in Multivariate Time Series through Granger Causal Discovery. In International Conference on Learning Representations

    2025

  13. [13]

    Han, X.; Zhang, L.; Wu, Y.; and Yuan, S. 2023. On Root Cause Localization and Anomaly Mitigation through Causal Inference. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, 699--708. ACM

    2023

  14. [14]

    Holm, S. 1979. A Simple Sequentially Rejective Multiple Test Procedure. Scandinavian Journal of Statistics, 6(2): 65--70

    1979

  15. [15]

    Huangfu, Q.; and Hall, J. A. J. 2018. Parallelizing the Dual Revised Simplex Method. Mathematical Programming Computation, 10(1): 119--142

    2018

  16. [16]

    Hussain, M.; Fidge, C.; Foo, E.; and Jadidi, Z. 2023. Discovering a Data Interpreted Petri Net Model of Industrial Control Systems for Anomaly Detection. Expert Systems with Applications, 230: 120511

    2023

  17. [17]

    Ikram, A.; Chakraborty, S.; Mitra, S.; Saini, S.; Bagchi, S.; and Kocaoglu, M. 2022. Root Cause Analysis of Failures in Microservices through Causal Discovery. In Advances in Neural Information Processing Systems, volume 35, 31158--31170

    2022

  18. [18]

    K.; Bagchi, S.; and Kocaoglu, M

    Ikram, A.; Lee, K.; Agarwal, S.; Saini, S. K.; Bagchi, S.; and Kocaoglu, M. 2025. Root Cause Analysis of Failures from Partial Causal Structures. In Chiappa, S.; and Magliacane, S., eds., Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence, volume 286 of Proceedings of Machine Learning Research, 1794--1818. PMLR

    2025

  19. [19]

    Jang, E.; Gu, S.; and Poole, B. 2017. Categorical Reparameterization with G umbel-Softmax. In International Conference on Learning Representations

    2017

  20. [20]

    Johansson, K. H. 2000. The Quadruple-Tank Process: A Multivariable Laboratory Process with an Adjustable Zero. IEEE Transactions on Control Systems Technology, 8(3): 456--465

    2000

  21. [21]

    Lee, S.; Purushwalkam Shiva Prakash, S.; Cogswell, M.; Ranjan, V.; Crandall, D.; and Batra, D. 2016. Stochastic Multiple Choice Learning for Training Diverse Deep Ensembles. In Advances in Neural Information Processing Systems, volume 29. Curran Associates, Inc

    2016

  22. [22]

    Li, M.; Li, Z.; Yin, K.; Nie, X.; Zhang, W.; Sui, K.; and Pei, D. 2022. Causal Inference-Based Root Cause Analysis for Online Service Systems with Intervention Recognition. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , 3230--3240. ACM

    2022

  23. [23]

    Li, Z.; Zhao, Y.; Han, J.; Su, Y.; Jiao, R.; Wen, X.; and Pei, D. 2021. Multivariate Time Series Anomaly Detection and Interpretation Using Hierarchical Inter-Metric and Temporal Embedding. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, 3220--3230

    2021

  24. [24]

    Lin, X.; Tagliapietra, N.; Li, K.; Halilaj, L.; and Luettin, J. 2026. StableRCA : Robust Graph-Agnostic Mechanism-Level Root Cause Analysis. arXiv:2606.05636

    Pith/arXiv arXiv 2026

  25. [25]

    Lohse, C.; Dhir, A.; Ba, A.; Eck, B.; Ruffini, M.; and Wahl, J. 2026. PRIM : Meta-Learned Bayesian Root Cause Analysis. arXiv:2605.08786

    Pith/arXiv arXiv 2026

  26. [26]

    Loshchilov, I.; and Hutter, F. 2019. Decoupled Weight Decay Regularization. In International Conference on Learning Representations. OpenReview.net

    2019

  27. [27]

    Lu, Z.; Geng, Z.; Li, W.; and Xie, M. 2026. Probability of Root Cause: A Counterfactual Definition and Its Identification. arXiv:2605.11776

    Pith/arXiv arXiv 2026

  28. [28]

    u ke, T.; D \

    Mehling, C. W.; Pieper, S.; L \"u ke, T.; D \"o belt, J.; and Ihlenfeldt, S. 2026. Enabling Joint Benchmarking of Automated Root Cause Analysis and Causal Discovery in Manufacturing Using the causRCA Dataset. Procedia CIRP, 139: 114--120

    2026

  29. [29]

    Nagalapatti, L.; Srivastava, A.; Sarawagi, S.; and Sharma, A. 2025. Robust Root Cause Diagnosis Using In-Distribution Interventions. In International Conference on Learning Representations

    2025

  30. [30]

    R.; Okati, N.; Garrido Mejia, S.; Bl \"o baum, P.; and Janzing, D

    Orchard, W. R.; Okati, N.; Garrido Mejia, S.; Bl \"o baum, P.; and Janzing, D. 2025. Root Cause Analysis of Outliers with Missing Structural Knowledge. In Advances in Neural Information Processing Systems, volume 38, 161634--161671

    2025

  31. [31]

    Perera, D.; Letzelter, V.; Mariotte, T.; Cort \'e s, A.; Chen, M.; Essid, S.; and Richard, G. 2024. Annealed Multiple Choice Learning: Overcoming Limitations of Winner-Takes-All with Annealing. In Advances in Neural Information Processing Systems, volume 37, 10541--10574

    2024

  32. [32]

    Pham, L.; Ha, H.; and Zhang, H. 2024. BARO : Robust Root Cause Analysis for Microservices via Multivariate Bayesian Online Change Point Detection. Proceedings of the ACM on Software Engineering, 1(FSE): 2214--2237

    2024

  33. [33]

    Suhr, H.; Kaltenpoth, D.; and Vreeken, J. 2026. Root Cause Analysis of Measurement and Mechanistic Anomalies. arXiv:2601.23026

    Pith/arXiv arXiv 2026

  34. [34]

    Umsonst, D.; Ruths, J.; and Sandberg, H. 2023. Finite Sample Guarantees for Quantile Estimation: An Application to Detector Threshold Tuning. IEEE Transactions on Control Systems Technology, 31(2): 921--928

    2023

  35. [35]

    Wang, M.; Zhou, D.; and Chen, M. 2023. Hybrid Variable Monitoring: An Unsupervised Process Monitoring Framework with Binary and Continuous Variables. Automatica, 147: 110670

    2023

  36. [36]

    S.; and Fr \"a nzle, M

    Weilbach, J.; Gerwinn, S.; Barsim, K. S.; and Fr \"a nzle, M. 2024. Counterfactual-Based Root Cause Analysis for Dynamical Systems. In Machine Learning and Knowledge Discovery in Databases. Research Track -- European Conference, ECML PKDD 2024, Proceedings, Part VI , volume 14946 of Lecture Notes in Computer Science, 303--319. Springer

    2024

  37. [37]

    Weng, X.; Xu, X.; Chang, L.; Hou, P.; Wang, G.; and Dustdar, S. 2022. Evidence Fusion-Based Alarm System Design Considering Coarse and Fine Changes of Process Variable. Journal of Process Control, 113: 68--79

    2022

  38. [38]

    K.; Devijver, E.; Gaussier, \'E .; and A \"i t-Bachir, A

    Zan, L.; Assaad, C. K.; Devijver, E.; Gaussier, \'E .; and A \"i t-Bachir, A. 2024. On the Fly Detection of Root Causes from Observed Data with Application to IT Systems. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, 5062--5069

    2024

  39. [39]

    2018 18th

    Wang, Ping and Xu, Jingmin and Ma, Meng and Lin, Weilan and Pan, Disheng and Wang, Yuan and Chen, Pengfei , title =. 2018 18th. 2018 , publisher =. doi:10.1109/CCGRID.2018.00076 , url =

    arXiv 2018

  40. [40]

    Proceedings of the 28th

    Li, Mingjie and Li, Zeyan and Yin, Kanglin and Nie, Xiaohui and Zhang, Wenchi and Sui, Kaixin and Pei, Dan , title =. Proceedings of the 28th. 2022 , publisher =. doi:10.1145/3534678.3539041 , url =

    arXiv 2022

  41. [41]

    Proceedings of the ACM on Software Engineering , volume =

    Pham, Luan and Ha, Huong and Zhang, Hongyu , title =. Proceedings of the ACM on Software Engineering , volume =. 2024 , month = jul, doi =

    2024

  42. [42]

    International Conference on Learning Representations , year =

    Han, Xiao and Absar, Saima and Zhang, Lu and Yuan, Shuhan , title =. International Conference on Learning Representations , year =

  43. [43]

    International Conference on Learning Representations , year =

    Nagalapatti, Lokesh and Srivastava, Ashutosh and Sarawagi, Sunita and Sharma, Amit , title =. International Conference on Learning Representations , year =

  44. [44]

    Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence , pages =

    Ikram, Azam and Lee, Kenneth and Agarwal, Shubham and Saini, Shiv Kumar and Bagchi, Saurabh and Kocaoglu, Murat , title =. Proceedings of the Forty-first Conference on Uncertainty in Artificial Intelligence , pages =. 2025 , editor =

    2025

  45. [45]

    Root Cause Analysis of Outliers with Missing Structural Knowledge , booktitle =

    Orchard, William Roy and Okati, Nastaran and Garrido Mejia, Sergio and Bl. Root Cause Analysis of Outliers with Missing Structural Knowledge , booktitle =. 2025 , url =

    2025

  46. [46]

    Enabling Joint Benchmarking of Automated Root Cause Analysis and Causal Discovery in Manufacturing Using the

    Mehling, Carl Willy and Pieper, Sven and L. Enabling Joint Benchmarking of Automated Root Cause Analysis and Causal Discovery in Manufacturing Using the. Procedia CIRP , volume =. 2026 , doi =

    2026

  47. [47]

    IEEE Transactions on Control Systems Technology , volume =

    Johansson, Karl Henrik , title =. IEEE Transactions on Control Systems Technology , volume =. 2000 , month = may, doi =

    2000

  48. [48]

    Scandinavian Journal of Statistics , volume =

    Holm, Sture , title =. Scandinavian Journal of Statistics , volume =. 1979 , doi =

    1979

  49. [49]

    Journal of the Royal Statistical Society: Series B (Methodological) , volume =

    Benjamini, Yoav and Hochberg, Yosef , title =. Journal of the Royal Statistical Society: Series B (Methodological) , volume =. 1995 , doi =

    1995

  50. [50]

    StableRCA: Robust Graph-Agnostic Mechanism-Level Root Cause Analysis

    Lin, Xiaoyu and Tagliapietra, Nicholas and Li, Kehan and Halilaj, Lavdim and Luettin, Juergen , title =. 2026 , eprint =. doi:10.48550/arXiv.2606.05636 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2606.05636 2026

  51. [51]

    Causal Structure-Based Root Cause Analysis of Outliers , booktitle =

    Budhathoki, Kailash and Minorics, Lenon and Bl. Causal Structure-Based Root Cause Analysis of Outliers , booktitle =. 2022 , editor =

    2022

  52. [52]

    Advances in Neural Information Processing Systems , volume =

    Ikram, Azam and Chakraborty, Sarthak and Mitra, Subrata and Saini, Shiv and Bagchi, Saurabh and Kocaoglu, Murat , title =. Advances in Neural Information Processing Systems , volume =. 2022 , doi =

    2022

  53. [53]

    Proceedings of the 32nd ACM International Conference on Information and Knowledge Management , pages =

    Han, Xiao and Zhang, Lu and Wu, Yongkai and Yuan, Shuhan , title =. Proceedings of the 32nd ACM International Conference on Information and Knowledge Management , pages =. 2023 , publisher =. doi:10.1145/3583780.3614995 , url =

    arXiv 2023

  54. [54]

    Advances in Neural Information Processing Systems , volume =

    Ji, Xiayan and Xue, Anton and Wong, Eric and Sokolsky, Oleg and Lee, Insup , title =. Advances in Neural Information Processing Systems , volume =. 2024 , doi =

    2024

  55. [55]

    Counterfactual-Based Root Cause Analysis for Dynamical Systems , booktitle =

    Weilbach, Juliane and Gerwinn, Sebastian and Barsim, Karim Said and Fr. Counterfactual-Based Root Cause Analysis for Dynamical Systems , booktitle =. 2024 , volume =. doi:10.1007/978-3-031-70365-2_18 , url =

    work page doi:10.1007/978-3-031-70365-2_18 2024

  56. [56]

    PRIM: Meta-Learned Bayesian Root Cause Analysis

    Lohse, Christopher and Dhir, Anish and Ba, Amadou and Eck, Bradley and Ruffini, Marco and Wahl, Jonas , title =. 2026 , eprint =. doi:10.48550/arXiv.2605.08786 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.08786 2026

  57. [57]

    Probability of Root Cause: A Counterfactual Definition and Its Identification

    Lu, Zitong and Geng, Zhi and Li, Wei and Xie, Min , title =. 2026 , eprint =. doi:10.48550/arXiv.2605.11776 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2605.11776 2026

  58. [58]

    Expert Systems with Applications , volume =

    Hussain, Mukhtar and Fidge, Colin and Foo, Ernest and Jadidi, Zahra , title =. Expert Systems with Applications , volume =. 2023 , doi =

    2023

  59. [59]

    ACM Computing Surveys , volume =

    Chandola, Varun and Banerjee, Arindam and Kumar, Vipin , title =. ACM Computing Surveys , volume =. 2009 , doi =

    2009

  60. [60]

    and Aubet, Fran

    Carmona, Chris U. and Aubet, Fran. Neural Contextual Anomaly Detection for Time Series , booktitle =. 2022 , doi =

    2022

  61. [61]

    Proceedings of the AAAI Conference on Artificial Intelligence , volume =

    Deng, Ailin and Hooi, Bryan , title =. Proceedings of the AAAI Conference on Artificial Intelligence , volume =. 2021 , doi =

    2021

  62. [62]

    Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =

    Li, Zhihan and Zhao, Youjian and Han, Jiaqi and Su, Ya and Jiao, Rui and Wen, Xidao and Pei, Dan , title =. Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages =. 2021 , doi =

    2021

  63. [63]

    and Ez-Zejjari, Imad and Zan, Lei , title =

    Assaad, Charles K. and Ez-Zejjari, Imad and Zan, Lei , title =. Proceedings of the 26th International Conference on Artificial Intelligence and Statistics , series =. 2023 , publisher =

    2023

  64. [64]

    and Devijver, Emilie and Gaussier,

    Zan, Lei and Assaad, Charles K. and Devijver, Emilie and Gaussier,. On the Fly Detection of Root Causes from Observed Data with Application to. Proceedings of the 33rd ACM International Conference on Information and Knowledge Management , pages =. 2024 , doi =

    2024

  65. [65]

    Root Cause Analysis of Measurement and Mechanistic Anomalies

    Suhr, Hendrik and Kaltenpoth, David and Vreeken, Jilles , title =. 2026 , eprint =. doi:10.48550/arXiv.2601.23026 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2601.23026 2026

  66. [66]

    NetCause: Counterfactual Learning for Root Cause Analysis in Large-Scale Networks

    Chraim, Fabien and Zhang, Jian and Janzing, Dominik and Song, Xiang and Faloutsos, Christos and Evans, John , title =. 2026 , eprint =. doi:10.48550/arXiv.2606.13543 , url =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2606.13543 2026

  67. [67]

    Zico and Koltun, Vladlen , title =

    Bai, Shaojie and Kolter, J. Zico and Koltun, Vladlen , title =. arXiv preprint arXiv:1803.01271 , year =. doi:10.48550/arXiv.1803.01271 , url =

    Pith/arXiv arXiv doi:10.48550/arxiv.1803.01271

  68. [68]

    Learning Phrase Representations Using

    Cho, Kyunghyun and van Merri. Learning Phrase Representations Using. Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing , pages =. 2014 , publisher =. doi:10.3115/v1/D14-1179 , url =

    doi:10.3115/v1/d14-1179 2014

  69. [69]

    IEEE Transactions on Control Systems Technology , volume =

    Umsonst, David and Ruths, Justin and Sandberg, Henrik , title =. IEEE Transactions on Control Systems Technology , volume =. 2023 , doi =

    2023

  70. [70]

    Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation , journal =

    Bengio, Yoshua and L. Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation , journal =. 2013 , doi =

    2013

  71. [71]

    International Conference on Learning Representations , year =

    Jang, Eric and Gu, Shixiang and Poole, Ben , title =. International Conference on Learning Representations , year =

  72. [72]

    Advances in Neural Information Processing Systems , volume =

    Lee, Stefan and Purushwalkam Shiva Prakash, Senthil and Cogswell, Michael and Ranjan, Viresh and Crandall, David and Batra, Dhruv , title =. Advances in Neural Information Processing Systems , volume =. 2016 , publisher =

    2016

  73. [73]

    Proceedings of the 42nd International Conference on Machine Learning , series =

    Cortes, Adrien and Rehm, Remi and Letzelter, Victor , title =. Proceedings of the 42nd International Conference on Machine Learning , series =. 2025 , publisher =

    2025

  74. [74]

    Annealed Multiple Choice Learning: Overcoming Limitations of Winner-Takes-All with Annealing , booktitle =

    Perera, David and Letzelter, Victor and Mariotte, Th. Annealed Multiple Choice Learning: Overcoming Limitations of Winner-Takes-All with Annealing , booktitle =. 2024 , doi =

    2024

  75. [75]

    Huangfu, Qi and Hall, J. A. J. , title =. Mathematical Programming Computation , volume =. 2018 , doi =

    2018

  76. [76]

    International Conference on Learning Representations , year =

    Loshchilov, Ilya and Hutter, Frank , title =. International Conference on Learning Representations , year =

  77. [77]

    Automatica , volume =

    Wang, Min and Zhou, Donghua and Chen, Maoyin , title =. Automatica , volume =. 2023 , doi =

    2023

  78. [78]

    Journal of Process Control , volume =

    Weng, Xu and Xu, Xiaobin and Chang, Leilei and Hou, Pingzhi and Wang, Guodong and Dustdar, Schahram , title =. Journal of Process Control , volume =. 2022 , doi =

    2022

This paper was first reviewed by deepseek-v4-flash on August 3, 2026.