You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, the exploration strategists can only traverse a space given by a predefined configuration. While this space may already be extremely large, it does not cover all the ways in which pipelines may be generated. For instance, there are cases when the user may exclude a component from a phase, thus artificially reducing the paths in the configuration space that may prove to be fruitful.
By exposing the Type Capabilities of components it would be possible to not only explore the configuration space from a set of predefined components, but also any such components that can be combined. In our efforts to obviate the manual work that comes in "discovering" pipelines, this is clearly the next obvious step.
What do we need to do to generalize our current execution model to enable this approach?
The text was updated successfully, but these errors were encountered:
Currently, the exploration strategists can only traverse a space given by a predefined configuration. While this space may already be extremely large, it does not cover all the ways in which pipelines may be generated. For instance, there are cases when the user may exclude a component from a phase, thus artificially reducing the paths in the configuration space that may prove to be fruitful.
By exposing the Type Capabilities of components it would be possible to not only explore the configuration space from a set of predefined components, but also any such components that can be combined. In our efforts to obviate the manual work that comes in "discovering" pipelines, this is clearly the next obvious step.
What do we need to do to generalize our current execution model to enable this approach?
The text was updated successfully, but these errors were encountered: