Google Introduces Neural Structured Learning In TensorFlow

NSL empowers TensorFlow users to effortlessly incorporate various structured signals for training neural networks.

Recently, Google announced Neural Structured Learning in TensorFlow, an easy-to-use open-source framework for training neural networks with structured signals.
 
Neural Structured Learning dubbed NSL can be used to build more efficient models for vision, language understanding, and prediction in general.
 
NSL framework implements Neural Graph Learning. It empowers you to train neural networks using graphs. The graphs can come from multiple sources such as Knowledge graphs, medical records, genomic data or multimodal relations.
 
Neural Structured Learning also generalizes to Adversarial Learning where the structure between input examples is dynamically constructed using adversarial perturbation.
 
It facilitates TensorFlow users to effortlessly incorporate different structured signals for training neural networks.
NSL
Source: Google 
 
In NSL, the structured signals(whether explicitly defined as a graph or implicitly learned as adversarial examples) are used to regularize the training of a neural network. This forces the model to learn accurate predictions by minimizing supervised loss, while maintaining the similarity among inputs from the same structure by minimizing the neighbor loss.