Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It was developed with a focus on enabling fast experimentations. Being able to go from ideas to results with the least possible delay is a key to doing good research.
This workshop is designed for those with entry-level deep learning knowledge who are interested in knowing more about the Keras package. We will go through basic Keras structures, demonstrate a simple image classifier and discuss the applications of deep learning in the insurance industry.
Additionally, in this workshop we will also be presenting a Natural Language Processing (NLP) deep learning task using Gensim Word2Vec jointly with Keras. In order to develop mathematical models on text corpus, it is good practice to convert text into a matrix representation. Gensim is such a clean open-source Python library to handle text data. It is a very robust and efficient tool that specializes in vector space and topic modeling.
Sample codes and sample data will be provided and demonstrated with Jupyter Notebook.
Dependencies (please install before the workshop):
Python >= 2.6. NumPy >= 1.3. SciPy >= 0.7 Jupyter Notebook Sklearn Keras Gensim