REVIEW 4 cited by
Disentangling Interactions and Dependencies in Feature Attribution
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
In explainable machine learning, global feature importance methods try to determine how much each individual feature contributes to predicting the target variable, resulting in one importance score for each feature. But often, predicting the target variable requires interactions between several features (such as in the XOR function), and features might have complex statistical dependencies that allow to partially replace one feature with another one. In commonly used feature importance scores these cooperative effects are conflated with the features' individual contributions, making them prone to misinterpretations. In this work, we derive DIP, a new mathematical decomposition of individual feature importance scores that disentangles three components: the standalone contribution and the contributions stemming from interactions and dependencies. We prove that the DIP decomposition is unique and show how it can be estimated in practice. Based on these results, we propose a new visualization of feature importance scores that clearly illustrates the different contributions.
Forward citations
Cited by 4 Pith papers
-
From Statistical to Structural Synergy: A Predictability Framework to Quantify the Effects due to High-Order Mechanisms
Structural synergy, defined as the excess predictive power of a joint model over the best additive model, isolates non-additive interaction mechanisms from dependency-driven statistical synergy in complex systems.
-
Information-theoretic Quantification of High-order Feature Effects in Classification Problems
A kNN-based CMI estimator applied to classification decomposes feature importance into unique, redundant, and synergistic components, validated on synthetic and real data.
-
Feature Attribution from First Principles
Under linearity and a continuity assumption, every feature attribution method is an integral against a signed measure, so the design choice reduces to picking the measure.
-
Saliency Maps are Ambiguous: Analysis of Logical Relations on First and Second Order Attributions
On synthetic AND/OR/XOR datasets with perfectly accurate models, every tested saliency method sometimes ranks a truly irrelevant input above a necessary one, so the scores cannot be trusted as relevance rankings.
Discussion (0). Continue with ORCID to comment.