Title

Workshop 3: Deep Learning with Python

Presenter Information/ Coauthors Information

David Zeng, Dakota State University

Presentation Type

Workshop

Abstract

Currently Deep Learning is required for at least 80% of jobs for data scientists and machine learning engineers. This workshop introduces Deep Learning concepts, models, and applications with Keras, the most popular high-level library for Deep Learning in Python. Topics include Keras’ sequential models, Convolutional Neural Networks for Image Classification, Recurrent Neural Networks (LSTM) for natural language processing. Some advanced and popular topics such as transfer learning with pre-trained models, reinforcement Q-learning with OpenAI Gym, and training Deep Learning models with GPU may be covered as well. While some experience in Python or Data Analytics may be beneficial, no previous knowledge about Deep Learning is required. All working documents (with tutorials and Python codes) will be provided.

Tentative outline of the workshop:

1. Introduction to the concepts of Deep Learning

2. Introduction to Keras Sequential models

3. Convolutional Neural Networks for image classification

4. Recurrent Neural Networks (LSTM) for language processing

5. (tentative) Introduction to Transfer Learning (There is a presentation on this by one of my students in the following day, so I wouldn't cover it in details)

6. (tentative) Introduction to Reinforcement Q-Learning with OpenAI Gym (depending on if there is time left)

Workshop 3 instructions

1. Bring laptops: No specific prerequisites are required.

2. Python 3 is needed. Installation via Anaconda distribution is recommended. Scikit-learn is needed. Jupyter notebook is needed.

3. Keras and Tensorflow are needed too. I do have a tutorial on this and so we can cover this at the beginning of the workshop for those who haven't done so yet. But I prefer they have the installations covered before the workshop.

Start Date

4-2-2019 1:00 PM

End Date

4-2-2019 5:00 PM

This document is currently not available here.

Share

COinS
 
Feb 4th, 1:00 PM Feb 4th, 5:00 PM

Workshop 3: Deep Learning with Python

Dakota Room 250 A/C

Currently Deep Learning is required for at least 80% of jobs for data scientists and machine learning engineers. This workshop introduces Deep Learning concepts, models, and applications with Keras, the most popular high-level library for Deep Learning in Python. Topics include Keras’ sequential models, Convolutional Neural Networks for Image Classification, Recurrent Neural Networks (LSTM) for natural language processing. Some advanced and popular topics such as transfer learning with pre-trained models, reinforcement Q-learning with OpenAI Gym, and training Deep Learning models with GPU may be covered as well. While some experience in Python or Data Analytics may be beneficial, no previous knowledge about Deep Learning is required. All working documents (with tutorials and Python codes) will be provided.

Tentative outline of the workshop:

1. Introduction to the concepts of Deep Learning

2. Introduction to Keras Sequential models

3. Convolutional Neural Networks for image classification

4. Recurrent Neural Networks (LSTM) for language processing

5. (tentative) Introduction to Transfer Learning (There is a presentation on this by one of my students in the following day, so I wouldn't cover it in details)

6. (tentative) Introduction to Reinforcement Q-Learning with OpenAI Gym (depending on if there is time left)

Workshop 3 instructions

1. Bring laptops: No specific prerequisites are required.

2. Python 3 is needed. Installation via Anaconda distribution is recommended. Scikit-learn is needed. Jupyter notebook is needed.

3. Keras and Tensorflow are needed too. I do have a tutorial on this and so we can cover this at the beginning of the workshop for those who haven't done so yet. But I prefer they have the installations covered before the workshop.