pith. sign in

arxiv: 2605.08688 · v1 · submitted 2026-05-09 · 💻 cs.AI · cs.DB· cs.LO

Reconciling Consistency-Based Diagnosis with Actual-Causality-Based Explanations

Leopoldo Bertossi This is my paper

Pith reviewed 2026-05-12 00:50 UTC · model grok-4.3

classification 💻 cs.AI cs.DBcs.LO
keywords consistency-based diagnosisactual causalitycausal responsibilityexplainable AIXAImodel-based diagnosiscausal explanations
0
0 comments X

The pith

Consistency-based diagnosis can be reconciled with actual causality to support explainable AI.

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

The paper sets out to connect consistency-based diagnosis, an approach that identifies faults by checking consistency with observed behavior, to the framework of actual causality and causal responsibility. It points out that the diagnosis community has developed tools for pinpointing minimal sets of components that explain inconsistencies, yet these have seen little use in explainable AI. A sympathetic reader would care because such links could let AI systems generate explanations that are both causally precise and diagnostically minimal, improving accountability in data management and decision systems.

Core claim

We establish, from the point of view of Explainable AI, connections between Consistency-Based Diagnosis on one side, and Actual Causality and Causal Responsibility on the other. CBD has received little attention from the XAI community. Connections between these two areas could have a fruitful impact on XAI and Explainable Data Management.

What carries the argument

The explicit mapping or reconciliation between consistency-based diagnosis principles and actual-causality explanations that allows diagnostic minimal sets to serve as causal accounts.

If this is right

  • Diagnostic consistency checks become available as a source of explanations inside XAI pipelines.
  • Causal responsibility measures can be computed using the same minimal hitting-set machinery already developed for diagnosis.
  • Explainable data management gains a new family of tools that treat data inconsistencies as causal events.
  • AI systems can produce explanations that are both minimal and causally grounded rather than purely statistical.

Where Pith is reading between the lines

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

  • Hybrid systems could run consistency-based diagnosis first to generate candidate explanations and then filter them with causal-responsibility scores.
  • Benchmarks from model-based diagnosis could be reused to evaluate the faithfulness of causal explanations generated by large language models.
  • The same reconciliation might extend to other forms of abductive reasoning in knowledge representation.

Load-bearing premise

That meaningful and fruitful connections between consistency-based diagnosis and actual causality can be established and will impact XAI and explainable data management.

What would settle it

A concrete diagnostic scenario in which every minimal consistency-based explanation fails to correspond to any actual cause or causal responsibility value under standard counterfactual definitions.

Figures

Figures reproduced from arXiv: 2605.08688 by Leopoldo Bertossi.

Figure 1
Figure 1. Figure 1: (a) A Binary Classifier. (b) Particular Input/Output. 2 [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (a) Successful Intervention. (b) Unsuccessful Intervention. As shown in [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: (a) Unsuccessful Intervention. (b) Successful Intervention. If, in addition to those two changes, we change again x2 to 1, we are now successful, as shown in [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Boolean Circuit. (b) Faulty Boolean Circuit. In this example, a logical model of the circuit, when it works properly, is a set of propositional formulas: {(x ←→ (a ∧ b)), (d ←→ (x ∨ c))}. However, our circuit at hand, by working abnormally, is not modeled by these formulas. Furthermore, the observation, Obs = {a, ¬b, c,¬d}, indicating that a and c are true, but b and d are false, is mutually inconsiste… view at source ↗
Figure 5
Figure 5. Figure 5: A Consistency Classifier As we saw in Example 3, with e = ⟨0, 0, 1, 0, 1, 0⟩, representing ⟨¬abA, ¬abO, a, , ¬b, c,¬d⟩, the classifier returns “no” (or 0). Then, we can apply actual causality by first declaring “features” abA, abO as endogenous, subject to inter￾ventions, while the others, a, b, c, d, are exogenous. They provide context and are not subject to interventions. By counterfactually intervening … view at source ↗
Figure 6
Figure 6. Figure 6: Causal Network Example 6. (ex. 4 cont.) The causal network in [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗
Figure 1
Figure 1. Figure 1: A deterministic and decomposable Boolean Circuit as a classifier. Figure [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
read the original abstract

We establish, from the point of view of Explainable AI (XAI), connections between Consistency-Based Diagnosis (CBD), on one side, and Actual Causality and Causal Responsibility, on the other. CBD has received little attention from the XAI community. Connections between these two areas could have a fruitful impact on XAI and Explainable Data Management.

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

0 major / 2 minor

Summary. The paper claims to establish connections between Consistency-Based Diagnosis (CBD) and Actual Causality/Causal Responsibility from an XAI perspective. It does so by constructing explicit translations between diagnostic models and structural causal models, under which minimal diagnoses correspond to actual causes under certain interventions. The central result is supported by parameter-free derivations and illustrative examples, with the goal of enabling fruitful cross-pollination between the two areas for XAI and explainable data management.

Significance. If the claimed translations and correspondences hold, the work could meaningfully advance XAI by bringing underutilized CBD techniques into causality-based explanation frameworks (and vice versa). Explicit strengths include the parameter-free derivations and the concrete examples that support the equivalence between minimal diagnoses and actual causes; these provide a solid formal foundation rather than mere assertions.

minor comments (2)
  1. The abstract is very concise and does not mention the key technical device (the explicit model translations); a one-sentence summary of the main construction would help readers decide whether to read further.
  2. Notation for the diagnostic models and the corresponding SCMs could be aligned more explicitly in the formal sections (e.g., by adding a small comparison table) to reduce cognitive load for readers who are expert in only one of the two literatures.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive summary, recognition of the paper's formal contributions, and recommendation for minor revision. We appreciate the acknowledgment of the potential impact on XAI through connections between consistency-based diagnosis and actual causality.

Circularity Check

0 steps flagged

No significant circularity detected in derivation chain

full rationale

The paper's central contribution consists of explicit translations between consistency-based diagnosis models and structural causal models, with minimal diagnoses shown to correspond to actual causes under specified interventions. These mappings are constructed from the independent definitions of the two frameworks and are parameter-free, with no fitted quantities, self-definitional reductions, or load-bearing self-citations that collapse the claimed equivalences back to the inputs by construction. The derivations remain self-contained against external benchmarks from diagnosis theory and actual causality literature.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no free parameters, axioms, or invented entities are specified or detectable.

pith-pipeline@v0.9.0 · 5339 in / 916 out tokens · 29246 ms · 2026-05-12T00:50:28.528020+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

26 extracted references · 26 canonical work pages · 1 internal anchor

  1. [1]

    and Szczepanski, N

    Audemard, G., Lagniez, J.-M., Marquis, P. and Szczepanski, N. On the Com- putation of Example-Based Abductive Explanations for Random Forests. Proc. IJCAI 2024, pp. 3679-3687

    work page 2024

  2. [2]

    and Monet, M

    Arenas, M., Barcelo, P., Bertossi, L. and Monet, M. On the Complexity of SHAP- Score-Based Explanations: Tractability via Knowledge Compilation and Non- Approximability Results.J. Mach. Learn. Res., 2023, 24(63):1-58

    work page 2023

  3. [3]

    Synthesis Lec- tures in Data Management

    Bertossi, L.Database Repairing and Consistent Query Answering. Synthesis Lec- tures in Data Management. Morgan & Claypool, 2011

    work page 2011

  4. [4]

    and Salimi, B

    Bertossi, L. and Salimi, B. From Causes for Database Queries to Repairs and Model-Based Diagnosis and Back.Theory of Computing Systems, 2017, 61(1):191- 232. 3 For responsibility under database integrity constraints, which goes along these lines, see [5, 8]. 12

    work page 2017

  5. [5]

    and Salimi, B

    Bertossi, L. and Salimi, B. Causes for Query Answers from Databases: Datalog Abduction, View-Updates, and Integrity Constraints.Int. J. Approximate Rea- soning, 2017, 90:226-252

    work page 2017

  6. [6]

    and Vagena, Z

    Bertossi, L., Li, J., Schleich, M., Suciu, D. and Vagena, Z. Causality-Based Ex- planation of Classification Outcomes. Proc. 4th WS Data Management for End- To-End Machine Learning (SIGMOD-DEEM), 2020, pp. 6:1-6:10

    work page 2020

  7. [7]

    Declarative Approaches to Counterfactual Explanations for Classifi- cation.Theory and Practice of Logic Programming, 2023, 23(3):559–593

    Bertossi, L. Declarative Approaches to Counterfactual Explanations for Classifi- cation.Theory and Practice of Logic Programming, 2023, 23(3):559–593

    work page 2023

  8. [8]

    From Database Repairs to Causality in Databases and Beyond

    Bertossi, L. From Database Repairs to Causality in Databases and Beyond. TLDKS, Springer LNCS 14160, 2023, pp. 119-131

    work page 2023

  9. [9]

    Attribution-Scores in Data Management and Explainable Machine Learning

    Bertossi, L. Attribution-Scores in Data Management and Explainable Machine Learning. Invited tutorial paper in Proc. ADBIS’23, Springer LNCS 13985, 2023, pp. 16-33

    work page 2023

  10. [10]

    and Le´ on, J

    Bertossi, L. and Le´ on, J. E. Efficient Computation of Shap Explanation Scores for Neural Network Classifiers via Knowledge Compilation. Proc. JELIA’23, Springer LNCS 14281, 2023, pp. 49-64

    work page 2023

  11. [11]

    and Pardal, N

    Bertossi, L. and Pardal, N. Sufficient Explanations in Databases and their Con- nections to Database Repairs. arXiv paper 2511.15623

    work page internal anchor Pith review arXiv

  12. [12]

    and Lafourcade, P

    Bienvenu, M., Figueira, D. and Lafourcade, P. Shapley Revisited: Tractable Re- sponsibility Measures for Query Answers. Proc. ACM on Management of Data, 2025, 3(2)112, pp. 1–26

    work page 2025

  13. [13]

    and Zick, Y

    Biradar, G., Izza, Y., Lobo, E., Viswanathan, V. and Zick, Y. Axiomatic Aggre- gations of Abductive Explanations. In Proc. AAAI 2024, 2024

    work page 2024

  14. [14]

    and Halpern, J

    Chockler, H. and Halpern, J. Responsibility and Blame: A Structural-Model Ap- proach.J. Artif. Intell. Res., 2004, 22:93-115

    work page 2004

  15. [15]

    Logic for Explainable AI

    Darwiche, A. Logic for Explainable AI. Proc. LICS 2023, pp. 1-11

    work page 2023

  16. [16]

    and Reiter, R

    de Kleer, J., Mackworth, A. and Reiter, R. Characterizing Diagnoses and Systems. Artif. Intell., 1992, 56(2-3):197-222

    work page 1992

  17. [17]

    and Pearl, J

    Halpern, J. and Pearl, J. Causes and Explanations: A Structural-Model Ap- proach. Part I: Causes.The British Journal for the Philosophy of Science, 2005, 56(4):843-887

    work page 2005

  18. [18]

    MIT Press, 2016

    Halpern, J.Actual Causality. MIT Press, 2016

    work page 2016

  19. [19]

    Huang, X., Cooper, M. C. Morgado, A., Planes, J. and Marques-Silva, J. Fea- ture Necessity and Relevancy in Machine Learning Explanations.J. Automated Reasoning, 2026, 70(1):4

    work page 2026

  20. [20]

    and Bl¨ obaum, P

    Janzing, D., Minorics, L. and Bl¨ obaum, P. Feature Relevance Quantification in Explainable AI: A Causal Problem. Proc. 23rd International Conference on Artificial Intelligence and Statistics (AISTATS) 2020, PMLR: Vol. 108

    work page 2020

  21. [21]

    Complexity of Consistent Query Answering in Databases under Cardinality-Based and Incremental Repair Semantics (extended version)

    Lopatenko, A. and Bertossi, L. Complexity of Consistent Query Answering in Databases under Cardinality-Based and Incremental Repair Semantics. Proc. ICDT, 2007, Springer LNCS 4353, pp. 179-193. Extended version in arXiv pa- per 1605.07159

    work page Pith review arXiv 2007

  22. [22]

    Meliou, A., Gatterbauer, W., Moore, K. F. and Suciu, D. The Complexity of Causality and Responsibility for Query Answers and Non-Answers. Proc. VLDB 2010, pp. 34-41

    work page 2010

  23. [23]

    Cambridge University Press, 2nd Ed., 2009

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

    work page 2009

  24. [24]

    A Theory of Diagnosis from First Principles.Artificial Intelligence, 1987, 32(1):57-95

    Reiter, R. A Theory of Diagnosis from First Principles.Artificial Intelligence, 1987, 32(1):57-95

    work page 1987

  25. [25]

    Causality for Machine Learning

    Sch¨ olkopf, B. Causality for Machine Learning. InProbabilistic and Causal Infer- ence: The Works of Judea Pearl, 2022, pp. 765-804. 13

    work page 2022

  26. [26]

    Model-Based Problem Solving

    Struss, P. Model-Based Problem Solving. InHandbook of Knowledge Represen- tation, Chap. 4, Elsevier, 2008, pp. 395-465. 14

    work page 2008