SGShift attributes concept shift to a sparse set of features by fitting a penalized generalized additive update term on top of the source model.
A survey of methods for explaining black box models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Explaining Concept Shift with Interpretable Feature Attribution
SGShift attributes concept shift to a sparse set of features by fitting a penalized generalized additive update term on top of the source model.