Improving the TMVA

Evolution Proposals, Apps & Ideas
1

Hello everyone,

I am working on improving the TMVA. There is a lot of work that needs to be taken care of like:

  1. Implementing new machine learning algorithms
  2. Adding more activation functions (hard-sigmoid, softplus, exponential linear unit,…)
  3. Adding more loss functions (cross entropy, hinge, …)
  4. Implementing L1L2 regularizer

If someone can guide me regarding this then it would be really helpful.

2

Hi,

We will be grateful for any contributions you are willing to make. Adding new algorithms is maybe more difficult, but things like adding more activation functions or improving the refularization for t he Deep learning module will be easier. Contributions can be made as pull requests on the ROOT github.
Please contact direct my self by email if you want to discuss this further

Best regards

Lorenzo