CSE and Amazon Workshop: Dive into Deep Learning Using MXNet

Introduction

Machine learning is the study of powerful techniques that can learn behavior from experience. This workshop brings in expertise from Amazon and will cover the fundamentals of machine learning, and focus in particular on deep learning, a powerful set of techniques driving innovations in areas as diverse as computer vision, natural language processing, and time-series analysis.

Event Details

Date: April 19-23, 2019 (Friday-Tuesday)
Time: 9:00 am - 12:00 nn (AM Session)
12:00 nn - 1:15 pm (Lunch Break)
1:15 pm - 2:45 pm (PM Session - 1st part)
2:45 pm - 3:15 pm (Break)
3:15 pm - 5:00 pm (PM Session - 2nd part)
Venue: LT-C, HKUST
Registration: Apply Online
Early Bird Registration Deadline: 31 March 2019 (Sunday)

Program

Date Topics to be Covered
19 April (Friday) Fundamentals:
  1. Introduction to deep learning
  2. Gluon Crash Course
  3. Write your first MLP
  4. Fine-tune your model
  5. Introduction to Convolutional Neural Networks
  6. Write your first CNN
  7. Find-tune your CNN
20 April (Saturday) Computer Vision:
  1. An introduction to Computer vision
  2. Computer vision use cases
  3. GluonCV
  4. Image data pre-processing
  5. Using ModelZoo for segmentation tasks
  6. Using ModelZoo for classification tasks
  7. Using ModelZoo for object Detection tasks
  8. Developing your own classification task
  9. Fine-tune your model
  10. Tips and tricks for large datasets
21 April (Sunday) Time-series analysis:
  1. Introduction to Autoregressive models
  2. Introduction to RNN
  3. RNN and LSTM Lab
  4. DeepAR + lab
  5. Remote Debugging
  6. LSTNet + Lab
22 April (Monday) NLP:
  1. GluonNLP
  2. Key Phrase Extraction
  3. MXBoard
  4. Sentiment Analysis in code-switching text + lab
  5. Intent detection and slot filling + lab
  6. NMT + Lab
23 April (Tuesday) MXNet:
  1. MMS (MXNet model server)
  2. TVM (MXNet model compiler)
  3. Multi GPU and cluster training
  4. Inference on the Edge
  5. Profiling
  6. Performance optimization tips and tricks
  7. MXNet and Keras

* Program will be conducted in English