-
Hi, in your paper seems no details about how dynamic diffuser works and how to get Influence Functions. can you provide more details? or the codes related to this part. Thank you
-
Add an implementation of "FastIF" (https://arxiv.org/abs/2012.15781) to the `interpret.influence_functions` module.
-
Hi, thanks for your nice work. I am planning to use influence functions in other settings and I have two question:
1. In general, do "helpful" data points with respect to a evaluation test point sh…
-
The setindex-functions for probability and utility matrices throw an error if a non-existent state is given, but all the functions have `probability_matrix.nodes[i]` when it should be `PM` or `UM`, de…
-
Implement Influence Functions to estimate correct standard errors for all decomposition components.
-
### Document to improve
Current description for the `pitch` parameter of sound related functions is as follow:
![image](https://github.com/Teardown-Issue-Tracker-Maintainers/Teardown-Issue-Tracker/a…
-
Hi,
I ran `PLIER` and `delayedPLIER` using the same parameters and the same SVD (calculated using `BiocSingular::runRandomSVD()` ).
The mean correlation of both Z matrices is around 0.8, and the…
-
### Abstract
- Use influence function to trace a model's prediction back to its training data.
- Approximation of influence function that requires gradients and Hessian vectors provides valuable i…
-
## 一言でいうと
DNNの判断根拠を理解するための試みで、ある学習データ(サンプル)がなかった場合のモデルへの影響を手がかりにする。通常だと該当サンプルを抜いての再学習が必要だが、該当サンプルのlossを増減させた場合の最適解を既存の最適解から導出するという技を使いこれをクリアしている。
### 論文リンク
https://arxiv.org/abs/1703.04730
…
-
[Understanding black-box predictions via influence functions](https://arxiv.org/abs/1703.04730)
How can we explain the predictions of a black-box model? In this paper, we use influence functions -- a…