Scientific evidence and limitations
BenchEWS distinguishes published mathematics, hypotheses, experimental indicators, software and external validation.
Peer-reviewed mathematical result
The AIP Advances article on spectral compression of stationary covariance matrices provides a peer-reviewed mathematical contribution. This does not validate every additional BenchEWS indicator or application.
Research hypotheses
Structural early-warning signals and the loss or recovery of self-correction are research questions. General predictive performance and causal interpretation require independent tests and domain-specific evidence.
Experimental indicators
Composite and exploratory indicators, including CRTI (EXPERIMENTAL_UNVALIDATED), are not universally validated risk scores or decision recommendations.
Software
BenchEWS Studio is an open-science computational tool. Reproducible computation or a functioning implementation is not equivalent to independent empirical validation.
Scope and limits
Transfer to organisations, societies or individuals is not established universally. No clinical, psychiatric or individual diagnostic use is claimed.
Independent reception
Website indexing, search-engine AI summaries and social-media references do not by themselves demonstrate independent scientific citation, replication or endorsement.
Primary peer-reviewed reference: AIP Advances · DOI: 10.1063/5.0337045