Advantage of a Conv1D layer in time series

What is the advantage of a Conv1D layer in time series? Could you explain me a bit more about the strengths of a Conv1D layer for time series and in which context it comes into play?

When I am not mistaken, the model of week 3 (excluding a Conv1D) performed better than the model including a Conv1D layer, on the synthetic data. Likewise for the sunspot activity, the difference appears comparatively small.

Please don’t let the number of dimensions change your opinion on Conv layers. It’d help if you could list the advantages of Conv2D layers (instead of using a Dense layer). Following the same train of thought should help you get to the case of Conv1D.