Charla

Deep Learning models in Python without gibberish

Maite Giménez

  • 24 September 2017, noon - 12:30 p.m.
  • Room Intelygenz
  • Idioma: en

Nowadays, a bunch of companies has become aware of all the information they can gather from what we write in social media. They can learn whether we liked or not their products, which are the demographic features of their users, etc. Except we write tons of things in social media, most of them have no value whatsoever, and they will need an army of people reading nonstop. But, do not despair, machine learning is here to help you automatize all these tasks.

Recently, we have seen a huge breakthrough in image processing [using deep learning techniques] [(https://www.cs.unc.edu/~wliu/papers/GoogLeNet.pdf]), and we computer linguistic begrudge what colleges can do in images.
So, if you want to use deep learning algorithms to predict the outcome of your system but you don't know exactly how to do it this is your talk.

If you are curious about machine learning and deep learning techniques in Python, this is your talk. This talk will have three parts. The first part will be mostly theoretical, and no background knowledge is required. However, some familiarity with machine learning slang will be helpful. This part will teach you about the algorithms used in deep learning, the more you understand the algorithms, the better you will apply them. The second part will be purely practical. You will see in action some of the deep learning methods for a classification task such as the prediction if a movie is going to pass the Bechtel test or not.
Later, I would explain why I chose the toolkit I use in my daily life and my insights of some mainstreams toolkits so you can pick whichever suits you best. 1. A gentle introduction to deep learning. (2') 2. Deep learning algorithms. (6') 3. Word-embeddings for deep learning. (4') 4. Predicting the Bechtel test. (8') 4. Deep learning toolkits (Keras, Theano & Tensorflow) and their applications. (5')