Psychology dealt with a very similar — maybe identical issue — with measurement and prediction in the 60s and 70s. The tension was between "empirically keyed" a

Psychology dealt with a very similar — maybe identical issue — with measurement and prediction in the 60s and 70s. The tension was between "empirically keyed" and "content" approaches to tests, the former being use of tests entirely based on large black box predictive item pools akin to large DL models, and tests based on some items selected based on some theory of test content. Many of the issues were similar. In the end I think they both lost out to an approach in which tests were empirically validated, but with items retained on their ability to meet various internal structural criteria. The "discovery" versus "justification" distinction reminds me of this in some ways. It would be akin to if some set of criteria were developed, not based on target prediction criteria, that DL model components would have to meet as constraints. Or, alternatively, you might formalize in some quantitative model what characterizes "justification" characteristics, in the sense of how to interpret a given DL model.