Interesting papers from NIPS 2016 II: f-GAN Training samplers with minimal divergence

The Intelligence of Information

I will review today one other interesting paper from the last NIPS 2016 Conference. This time the paper that I’ve chosen is about a relatively new unsupervised  machine learning method implemented by a system of two neural networks called Generative Adversarial Networks (GAN). This is a technique that is being used with some success, specially in Computer Vision applications, but the authors of this paper claim possible applications within natural language processing and signal processing frameworks.

One aspect of this method that is somehow innovative is the interplay of two neural networks which are performing two opposed tasks in an adversarial zero-sum game setting. One would have thought this to be unproductive, but we should recall that this is only a methodological approach; this is intended to improve how training data in a machine learning framework helps the generation process of real test data in order for the respective algorithm…

View original post 2,162 more words


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s