Stability theory of game-theoretic group feature explanations for machine learning models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2021
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909282972205056 |
|---|---|
| author | Miroshnikov, Alexey Kotsiopoulos, Konstandinos Filom, Khashayar Kannan, Arjun Ravi |
| author_facet | Miroshnikov, Alexey Kotsiopoulos, Konstandinos Filom, Khashayar Kannan, Arjun Ravi |
| contents | In this article, we study feature attributions of Machine Learning (ML) models originating from linear game values and coalitional values defined as operators on appropriate functional spaces. The main focus is on random games based on the conditional and marginal expectations. The first part of our work formulates a stability theory for these explanation operators by establishing certain bounds for both marginal and conditional explanations. The differences between the two games are then elucidated, such as showing that the marginal explanations can become discontinuous on some naturally-designed domains, while the conditional explanations remain stable. In the second part of our work, group explanation methodologies are devised based on game values with coalition structure, where the features are grouped based on dependencies. We show analytically that grouping features this way has a stabilizing effect on the marginal operator on both group and individual levels, and allows for the unification of marginal and conditional explanations. Our results are verified in a number of numerical experiments where an information-theoretic measure of dependence is used for grouping. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2102_10878 |
| institution | arXiv |
| publishDate | 2021 |
| record_format | arxiv |
| spellingShingle | Stability theory of game-theoretic group feature explanations for machine learning models Miroshnikov, Alexey Kotsiopoulos, Konstandinos Filom, Khashayar Kannan, Arjun Ravi Computer Science and Game Theory Probability 91A06, 91A12, 91A80, 46N30, 46N99, 68T01 In this article, we study feature attributions of Machine Learning (ML) models originating from linear game values and coalitional values defined as operators on appropriate functional spaces. The main focus is on random games based on the conditional and marginal expectations. The first part of our work formulates a stability theory for these explanation operators by establishing certain bounds for both marginal and conditional explanations. The differences between the two games are then elucidated, such as showing that the marginal explanations can become discontinuous on some naturally-designed domains, while the conditional explanations remain stable. In the second part of our work, group explanation methodologies are devised based on game values with coalition structure, where the features are grouped based on dependencies. We show analytically that grouping features this way has a stabilizing effect on the marginal operator on both group and individual levels, and allows for the unification of marginal and conditional explanations. Our results are verified in a number of numerical experiments where an information-theoretic measure of dependence is used for grouping. |
| title | Stability theory of game-theoretic group feature explanations for machine learning models |
| topic | Computer Science and Game Theory Probability 91A06, 91A12, 91A80, 46N30, 46N99, 68T01 |
| url | https://arxiv.org/abs/2102.10878 |