Deep Autoencoder in RapidMiner

GottfriedGottfried Member Posts: 17 image Maven
in Product Ideas
I crave for having an autoencoding neural network for advanced non-linear dimensionality réduction in RapidMiner. The only operator close to it today is the Replicator Neural Network (RNN), in the Anomaly Detection 3rd party extension. If you share my request, let it be known!  ;) 
Tagged:
15
Up
15 votes

Open for Voting · Last Updated

Done and open for voting cc @varunm1 @Gottfried PROD-228

Comments

  • varunm1varunm1 Member Posts: 1,207 image Unicorn
    @sgenzer can you have this in ideas section? We are desperate to vote on this  :smiley:
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

  • WTWT Member Posts: 5 image Contributor II
    Second that  - be good to have this native please
  • ZKuiperZKuiper Member Posts: 11 image Contributor II
    This is also something I have stumbled upon wishing I had.
  • hbajpaihbajpai Member Posts: 102 image Unicorn
    This would be great to have for sure!
    Best,
    Harshit
  • ROConnellROConnell Member Posts: 1 image Learner II
    I agree, this would be a great asset!
  • jacobcybulskijacobcybulski Member, University Professor Posts: 391 image Unicorn
    I suspect the current Deep Learning extension can generate an Autoencoder, however, I cannot see how to build a Variational Autoencoder in RapidMiner. Also, if we are to get an Autoencoder, why not taking a step further and build the whole generative model, such as GAN? (if you can build GANs, Autoencoders come for free). 
  • varunm1varunm1 Member Posts: 1,207 image Unicorn
    @jacobcybulski

    That is true, GAN can help in both ways. I discussed with @pschlunder and currently there aren't any specifics available, might come in future.
    Regards,
    Varun
    https://www.varunmandalapu.com/

    Be Safe. Follow precautions and Maintain Social Distancing

  • mortizmortiz Member Posts: 20 image Maven
    This would be a fantastic addition
Sign In or Register to comment.