Does intellij or jetbarins in general uses Machine...
# randomk
kartikpatodi
06/26/2018, 5:58 PMDoes intellij or jetbarins in general uses Machine Learning in any of their products?
👌 1
k
karelpeeters
06/27/2018, 9:27 AMFor what @orangy?
o
orangy
06/27/2018, 9:48 AM
For various things 🙂
🧌 6
🤔 1
l
louiscad
06/27/2018, 10:08 AM@orangy Can we expect some machine learning Kotlin goodness from JetBrains before 2020?
louiscad
06/27/2018, 10:08 AMI mean, which we can use ourselves, without being forced to use python to get started
o
orangy
06/27/2018, 10:08 AM
What’s wrong with projects by @thomasnield et al in #C4W52CFEZ ?
l
louiscad
06/27/2018, 10:19 AMNot much 🙂
The few things:
- Datalore looks to be a great learning environment, but zero Kotlin there.
- Resources I see in #C4W52CFEZ are usually JVM only. I'd like multiplatform Kotlin support (but I agree this is not very important at this moment, JVM support is already fine)
- Some close to real life examples, in Kotlin from JetBrains could be nice.
I heard you used machine learning for autocomplete suggestions in IDEs. Is this open source, and in Kotlin? If yes, I'd love to see a blog post explaining the approach so newbies in datascience can grasp the thing, and get started.
k
kartikpatodi
06/27/2018, 11:26 AM@orangy various things like
o
orangy
06/27/2018, 11:31 AM
Like processing big chunks of unstructured data and training models 🙂 Sorry, can’t say anything more specific.
⛔ 1
🙊 1
k
kartikpatodi
06/27/2018, 11:33 AMCool
t
thomasnield
06/27/2018, 9:56 PMYou got to keep in mind Machine Learning is not so generalized that you can easily create an XY Whizbanger library that solves all ML and optimization problems. You can generalize a specific approach (e.g. Naive Bayes or Neural Networks) or build for a specific task. But building a comprehensive and agreeable ML suite library for production (and not just exploration) is a sizable effort.
I've also found focusing on one methodology and building a library around that is going to be more agreeable as a dependency and a development effort for those dogfooding it.
There are many libs (even beyond ML) that'd be nice to see in multiplatform, but that may take time. For now, there are MANY great libraries already available in Java. For instance, check out SMILE.
👍 1
2 Views