Machine Learning - Neural Networks

Tensor Flow - Machine learning, ML tools that we use at Woonkly for data processing in the social network

Woonkly uses Tensor Flow tool which was named this way because it works by the flow of tensors in the form of a computational graph. Tensors are basically multidimensional arrays.

A computer graph has a network of nodes, and each node performs an operation such as addition, multiplication, or evaluation of some multivariate equation. The nodes represent mathematical operations and the edges represent tensors. TensorFlow can be used to build any kind of deep learning algorithms like:

  • CNN

  • RNN

  • DBN

  • FeedForward Neural Network

  • Algorithms for natural language processing

  • Others

There are various programming elements in TensorFlow such as constants, variables, placeholders, sessions, etc. Each one has its own functionalities and is used to build any Deep Learning model.

Tensoflow uses Python to provide a convenient front-end API for building applications with the framework, while running high-performance C ++ applications. Woonkly uses TensorFlow, that way it can "learn" and run deep "Neural Networks" to:

  • Classify handwritten texts

  • Do image recognition

  • Embedded words

  • Recurrent neural networks

  • Sequence-to-sequence models for machine translation

  • Natural language processing

  • Simulations based on PDE (partial differential equation).

You can test how the algorithm works here: