HLearn: A Machine Learning Library for Haskell (2013) [pdf]
1–10 of 23 posts
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#2a newer library targeted at deep learning: https://github.com/HuwCampbell/grenade
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#3What are the advantages of Haskell at deep learning, given that
1. Graph structures are notoriously difficult to model in functional languages.
2. The software-engineering side of deep learning is not all that difficult (e.g. using Keras is quite simple).
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#4I guess people should stop developing new libraries that already exist for other languages (Re-inventing the wheel?).
Even if the library is novel, people should write it for languages that are popular.
We should just stick to the stack we have today and hope that someday, AI-powered tools can rewrite all the libraries for us in our favorite language. Till then, people should keep working on the more important stuff than maintaining libraries for languages like Haskell, Clojure, etc.
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#5What are the advantages of Haskell at deep learning, given that 1. Graph structures are notoriously difficult to model in functional languages. 2. The software-engineering side of deep learning is not all that difficult (e.g. using Keras is quite simple).
You can for example prevent "runtime" bugs due to combining layers/vectors/matrixes of the wrong dimensions.
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#6What are the advantages of Haskell at deep learning, given that 1. Graph structures are notoriously difficult to model in functional languages. 2. The software-engineering side of deep learning is not all that difficult (e.g. using Keras is quite simple).
You can for example prevent "runtime" bugs due to combining layers/vectors/matrixes of the wrong dimensions.
regarding compiling
>Grenade layers are normal haskell data types which are an instance of Layer, so it's easy to build one's own downstream code. We do however provide a decent set of layers, including convolution, deconvolution, pooling, pad, crop, logit, relu, elu, tanh, and fully connected.
it's called a README for a reason.
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#7What are the advantages of Haskell at deep learning, given that 1. Graph structures are notoriously difficult to model in functional languages. 2. The software-engineering side of deep learning is not all that difficult (e.g. using Keras is quite simple).
Also, correct me if I'm wrong, but the state of the art seems to be happening in Tensorflow/Keras; so committing to a different platform could mean you are systematically lagging behind in this new field.
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#8Hi everyone, author of HLearn here :)
This is a bit awkward for me as I've paused the development of HLearn and emphatically do not recommend anyone use it. The main problem is that Haskell (which I otherwise love) has poor support for numerical computing. I've tried developing an alternative standard library to improve the situation (https://github.com/mikeizbicki/subhask), but Haskell's type system isn't yet powerful enough to do what I want. I'm sure the type system will have the needed features in 5-10 years, and I'd rather wait and do it right.
If you have any questions, I'd be happy to answer them.
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#9What are the advantages of Haskell at deep learning, given that 1. Graph structures are notoriously difficult to model in functional languages. 2. The software-engineering side of deep learning is not all that difficult (e.g. using Keras is quite simple).
We have to remember that machine learning is much more than just ANN, or, for that sake, matrix manipulation:
Re: HLearn: A Machine Learning Library for Haskell (2013) [pdf]
#10Hi everyone, author of HLearn here :) This is a bit awkward for me as I've paused the development of HLearn and emphatically do not recommend anyone use it. The main problem is that Haskell (which I otherwise love) has poor support for numerical computing. I've tried developing an alternative standard library to improve the situation ( https://github.com/mikeizbicki/subhask ), but Haskell's type system isn't yet powe…
What are you using instead?