Update 07-autodiff.md

1 files changed, 1 insertions, 1 deletions

diff --git a/docs/user-guide/07-autodiff.md b/docs/user-guide/07-autodiff.md
index 0751d4296..a2a06b64b 100644
--- a/docs/user-guide/07-autodiff.md
+++ b/docs/user-guide/07-autodiff.md
@@ -127,7 +127,7 @@ Where $$\omega' \in \mathbf{w}'$$ represents the partial derivative of $$\omega_
 Given this definition, $$\mathbb{F}[f]$$ can be used as a forward propagation function that is able to compute $$\frac{\partial f_i}{\partial \omega_0}$$ from $$\frac{\partial \omega_{i-1}}{\partial \omega_0}$$.
 
 ### Backward Propagation of Derivatives
-When the backpropagation algorithm to train a neural network, we are more interested in figuring out the partial derivative of the final system output with regard to a parameter $$\omega_i$$ in $$f_i$$. To do so, we generally utilize the backward derivative propagation function
+When using the backpropagation algorithm to train a neural network, we are more interested in figuring out the partial derivative of the final system output with regard to a parameter $$\omega_i$$ in $$f_i$$. To do so, we generally utilize the backward derivative propagation function
 
 $$\mathbb{B}[f_i] = f_i^{-1}(\frac{\partial Y}{\partial f_i}) = \frac{\partial Y}{\partial \mathbf{w}_i}$$