[FIX] Typo in notebook.

Author: ctallec

main_regression.ipynb

@@ -48,7 +48,7 @@
    "metadata": {},
    "source": [
     "# Building the model\n",
-    "The model used is a two layer MLP. The particularity here is that for every input $x$, the model outputs both a mean prediction $\\mu(x)$ and a standard deviation $\\sigma(x) = \\log(1 + e^\\rho(x))$. $\\sigma$ is parameterized this way to avoid negative values."
+    "The model used is a two layer MLP. The particularity here is that for every input $x$, the model outputs both a mean prediction $\\mu(x)$ and a standard deviation $\\sigma(x) = \\log(1 + e^{\\rho(x)})$. $\\sigma$ is parameterized this way to avoid negative values."
    ]
   },
   {