Weakening statistical assumptions to survive distribution shift — is not yet combined with → Formal verification and control theory recur as safety building blocks

The paper treats detecting distributional shift and responding to it as two separate problems, and splits the response tools out into formal-methods-imported. Covariate shift, partially specified models, unsupervised risk estimation, and training on multiple distributions are all ways of detecting or bounding confidence around a shift once it may have occurred. Reachability analysis and robust policy improvement, both formal-methods-imported members, are introduced immediately afterward as candidate responses, falling back to a conservative baseline policy once out-of-distribution operation is suspected. The paper flags the seam explicitly: to its authors' knowledge, this control-theoretic work has "not yet been combined with methods for detecting out-of-distribution failures of a model" (concrete-problems, §"How to respond when out-of-distribution.", p. 18). Neither theme's own narrative registers this as unfinished; each describes its half of the pipeline, detection or response, as though it were the whole story.