search

whitead / dmol-book

Deep learning for molecules and materials book
https://dmol.pub
Other
618 stars 121 forks source link

Typo #175

Closed jasperhyp closed 2 years ago

jasperhyp commented 2 years ago

https://dmol.pub/dl/data.html#input-data-equivariances

RHS in formula (9.2) should be $\mathbf{XW}+b-10|\mathbf W|$

Also, I wonder what $\mathbf W$ is like in this case. $\hat{E}$ is a scalar and $\mathbf X$ is a matrix, so there seems to be an aggregation function put upon $\mathbf{XW}$. Assuming $\mathbf W$ is some message passing with aggregation, claiming it to have "permutation variance" might not hold, as long as the aggregation is permutation invariant. Would appreciate it if anyone could clarify this!

whitead commented 2 years ago

There were a lot of problems with that example. Thanks for pointing it out. Fixed (I think!)