MétaCan
Menu
Back to cohort
Record W4414829247 · doi:10.4230/lipics.csl.2026.25

Constructing Witnesses for Lower Bounds on Behavioural Distances

2025· article· en· W4414829247 on OpenAlex
Ruben Turkenburg, Harsh Beohar, Franck van Breugel, Clemens Kupke, Jurriaan Rot

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

fundA Canadian funder is recorded on the work.
no affNo Canadian affiliation: this work is invisible to an affiliation-only frame.
No Canadian affiliation. An affiliation-only frame, the usual design, would never have seen this work. It is one of the works that make the case for inverting the frame.

Bibliographic record

VenueLeibniz-Zentrum für Informatik (Schloss Dagstuhl) · 2025
Typearticle
Languageen
FieldComputer Science
TopicAdversarial Robustness in Machine Learning
Canadian institutionsnot available
FundersEngineering and Physical Sciences Research CouncilNatural Sciences and Engineering Research Council of CanadaNederlandse Organisatie voor Wetenschappelijk OnderzoekLeverhulme Trust
KeywordsSoundnessMathematical proofBounding overwatchCompleteness (order theory)Probabilistic logicEquivalence (formal languages)ConstructiveContext (archaeology)Modal logic

Abstract

fetched live from OpenAlex

Behavioural distances provide a robust alternative to notions of equivalence such as bisimilarity in the context of probabilistic transition systems. They can be defined as least fixed points, whose universal property allows us to exhibit upper bounds on the distance between states, showing them to be at most some distance apart. In this paper, we instead consider the problem of bounding distances from below, showing states to be at least some distance apart. Contrary to upper bounds, it is possible to reason about lower bounds inductively. We exploit this by giving an inductive derivation system for lower bounds on an existing definition of behavioural distance for labelled Markov chains. This is inspired by recent work on apartness as an inductive counterpart to bisimilarity. Proofs in our system will be shown to closely match the behavioural distance by soundness and (approximate) completeness results. We further provide a constructive correspondence between our derivation system and formulas in a modal logic with quantitative semantics. This logic was used in recent work of Rady and van Breugel to construct evidence for lower bounds on behavioural distances. Our constructions provide smaller witnessing formulas in many examples.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.001
metaresearch head score (Gemma)0.001
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesMeta-epidemiology (narrow), Scholarly communication
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Theoretical or conceptual · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: none
Teacher disagreement score0.846
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0010.001
Meta-epidemiology (narrow)0.0010.000
Meta-epidemiology (broad)0.0010.000
Bibliometrics0.0000.001
Science and technology studies0.0010.000
Scholarly communication0.0010.003
Open science0.0020.001
Research integrity0.0000.001
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.014
GPT teacher head0.301
Teacher spread0.287 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it