How can we improve spark-rapids-ml?

[eordentlich]

We welcome suggestions on any improvements, and specifically input on the following:

• Which additional Spark MLlib algorithms should be GPU accelerated and included in spark-rapids-ml?
• What missing functionality for already included accelerated ML algorithms should be implemented?
• What GPU accelerated ML libraries/implementations would you like to be able to run in Spark clusters?
• If you like what we already have, what are some barriers to adoption if any?