From 03c14dae66af70883bc1dd0b10a83509daa76ff6 Mon Sep 17 00:00:00 2001
From: Nicolas GUILLARD <151741326+nicolasguillard@users.noreply.github.com>
Date: Fri, 13 Sep 2024 22:55:35 +0200
Subject: [PATCH] BugFix #402

Update 06.5-example-based-influence-fct.Rmd updating the latex code of the equation of miminizer for loss with upweighted point z is wrong in section 10.5.2
---
 manuscript/06.5-example-based-influence-fct.Rmd | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/manuscript/06.5-example-based-influence-fct.Rmd b/manuscript/06.5-example-based-influence-fct.Rmd
index cb154bc9d..a7d139dd5 100644
--- a/manuscript/06.5-example-based-influence-fct.Rmd
+++ b/manuscript/06.5-example-based-influence-fct.Rmd
@@ -340,7 +340,7 @@ The following section explains the intuition and math behind influence functions
 
 The key idea behind influence functions is to upweight the loss of a training instance by an infinitesimally small step $\epsilon$, which results in new model parameters:
 
-$$\hat{\theta}_{\epsilon,z}=\arg\min_{\theta{}\in\Theta}(1-\epsilon)\frac{1}{n}\sum_{i=1}^n{}L(z_i,\theta)+\epsilon{}L(z,\theta)$$
+$$\hat{\theta}_{\epsilon,z}=\arg\min_{\theta{}\in\Theta}\frac{1}{n}\sum_{i=1}^n{}L(z_i,\theta)+\epsilon{}L(z,\theta)$$
 
 where $\theta$ is the model parameter vector and $\hat{\theta}_{\epsilon,z}$ is the parameter vector after upweighting z by a very small number $\epsilon$.
 L is the loss function with which the model was trained, $z_i$ is the training data and z is the training instance which we want to upweight to simulate its removal.