Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead.¶
Rudin, C. (2019). Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead. Nature Machine Intelligence, 1, 206-215.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Rashomon Effect
- The discipline the prime imposes follows directly: a claim like "the model learned that feature X drives the outcome" is not licensed by the data, because another equally accurate model in the Rashomon set ignores X entirely.
This sourceArgues that with a Rashomon set of equally accurate models, a claim that the model 'learned' a given feature is not licensed by the data alone.
Supported in partVerified against the work's full text
Rudin defines the Rashomon set and says variable importance in one model of a set of near-equal models carries no causal weight, but never states the feature-X-learned rule.
“The importance of the variables in the model does not reflect a causal relationship between the variables and the outcomes.”
- The discipline the prime imposes follows directly: a claim like "the model learned that feature X drives the outcome" is not licensed by the data, because another equally accurate model in the Rashomon set ignores X entirely.
Verification¶
Does it exist? Confirmed. This work's DOI resolves to a registered record, which fixes its identity. That is all it fixes.
Does it back the claim? Read against the text for 1 of 1 citation: 1 supported in part. Each verdict is shown under its citation below, with what in the work backs the sentence.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:7584a7a0d5ba · see in the full table