Skip to content

Equivariant Neural Network Chapter #245

@whitead

Description

@whitead

I would have thought based on the earlier section’s notation that moving a point to (-4,3) should correspond to t_-4,3, but here it’s written t_-3,4 and repeated. Can you clarify?

Then later in the section where you define G-function transforms, it seems that if you consider the one dimensional case where g = t_10, if you chose x = 0 at the origin, then gx = 10 and (g^-1)x = -10, so by your definition f’(0) = f(-10). But then this new f’ would be a right-shifted copy of f (i.e. each point in f’ would be 10 units to the right of its coordinate in f), while the section describes it as a left-shift. Can you clarify this?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions