Deep learning system can be viewed as conceptually comprising a deep learning module and an application module as shown in the figure above.  The deep learning module accepts training data as input and produces a model (by tuning millions/billions of parameters) as output, while the application module accepts new input and applies the model produced by the deep learning module to generate the outcome.  Our current deep learning applications include, but not limited to:

• Machine Translation
• Study of Entrepreneurial Spirit Leading to Better Stock Return and Financial Performance
• Detection of Hong Kong English in Students’ Writing and Good Practice Analysis
• Detection of Fake News in Online Media
• Intelligent Configurations (iCON) Design
• AI Smart Retail
• Deep Learning Applications on Weather (Object Recognition and Interpretation)
• … and more coming!

 Form for project collaboration with the DLC

• Project Collaboration Proposal – Vetting Form
• Example of Project Collaboration Proposal – Vetting Form