A Reference-Relative Preorder on Causal Transformations: Relation to Empowerment and Controllabilit

We propose a reference-relative preorder on causal transformations X_{t+1}=F(X_t,E_t), comparing systems by the reference-outcomes that interventions on their exogenous input can reliably bring about. We show that bare reachability rewards noise, introduce a reliability-weighted refinement, and relate the order to empowerment: exactly in the deterministic case, via a two-sided Fano-type bound in general, and with a negative result on the granularity of point-outcome reliability. For linear systems the order reduces to controllable-subspace inclusion, the Loewner order on Gramians, and is refined by capacity in the linear-Gaussian case. We make no claim of a universal scalar measure of intelligence; derivations of the intervention space U(F) and of the reference R remain open.v

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24333445
Primary Topic
Reinforcement Learning in Robotics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

A Reference-Relative Preorder on Causal Transformations: Relation to Empowerment and Controllabilit

Chandra Mouli Krishna Moorthy
Zenodo (CERN European Organization for Nuclear Research)
Reinforcement Learning in Robotics
article

A Reference-Relative Preorder on Causal Transformations: Relation to Empowerment and Controllabilit

Chandra Mouli Krishna Moorthy
article en

Abstract

We propose a reference-relative preorder on causal transformations X_{t+1}=F(X_t,E_t), comparing systems by the reference-outcomes that interventions on their exogenous input can reliably bring about. We show that bare reachability rewards noise, introduce a reliability-weighted refinement, and relate the order to empowerment: exactly in the deterministic case, via a two-sided Fano-type bound in general, and with a negative result on the granularity of point-outcome reliability. For linear systems the order reduces to controllable-subspace inclusion, the Loewner order on Gramians, and is refined by capacity in the linear-Gaussian case. We make no claim of a universal scalar measure of intelligence; derivations of the intervention space U(F) and of the reference R remain open.v

Zenodo (CERN European Organization for Nuclear Research)
Reinforcement Learning in Robotics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon