Deep Neural Networks, the Graphics Processing Unit (GPU) and Big Data were under the spotlight during the Deep Learning Fundamentals for Computer Vision workshop held by UKZN’s School of Management, IT and Governance.
The workshop was facilitated by an Associate Professor in High Performance Computing at Huddersfield University in England, Dr Violeta Holmes, and the School’s Information Systems and Technology lecturer, Dr Upasana Singh, in collaboration with the NVIDIA Deep Learning Institute which trains developers, data scientists, researchers, and students on practical experience and certification on artificial intelligence and accelerated computing to solve real-world problems.
‘We all hear about how artificial intelligence is taking over the world. Deep learning is a new field that is applied in different areas of our lives showing that there are many more applications of deep learning,’ said Holmes.
The workshop trained participants on how to remove constraints associated with traditional programming through the following:
• Implementation of common deep learning workflows, such as image classification and object detection
• Experimentation with data, training parameters, network structure, and other strategies to increase performance and capability of neural networks
• Integration and deployment of neural networks in applications to start solving sophisticated real-world problems
‘Today you will be using only a browser to explore the world of deep learning,’ said Holmes. ‘We will demonstrate how to “deploy” solutions, created with Deep Learning, into real-world applications, highlighting the difference between training environments and deployment environments, and outline how to create applications that use the state-of-the-art in Deep Learning with existing models,’ explained Holmes.
After the workshop participants were given 30 days to complete any outstanding work from the Fundamentals of Deep Learning for Computer Vision course using online resources to receive a NVIDIA DLI Certification to recognise subject matter competency and support professional career growth.
Words: Thandiwe Jumo