Mikio L. Braun

ML and AI expert

Back to archive

Thread

2 tweets

1
I see that some thinking also went into how to evolve your tests to guide the development. Very interesting subject! In DS, there is something similar going from subsets and toy data to increasingly complex and larger data sets.
1
2
@lomin While I was researcher, especially the use of parameterized toy data (eg different noise levels) was a very important part of the work. I see too little of it in practice.