The application of Deep Learning to biometrics has proven to be extremely successful. Aurora is now offering “Deep Learning” as a service, to create bespoke AI solutions for other industries. AI has huge potential to improve traditional machine vision solutions and offers alternatives to manual analysis of data in many sectors. A brief overview follows…
What is Artificial Intelligence?
Artificial Intelligence (AI) is a computer program, typically incorporating an artificial neural network, that can learn to perform complex intellectual tasks. It comes in two forms:
- Strong AI (or General AI): A hypothetical system capable of performing any intellectual task, similar to a human. This is the end-goal of many AI researchers, but it is not what we do at Aurora.
- Narrow AI (or Applied AI): An Artificial Intelligence created to solve one specific problem. Once created, it cannot be applied to another task. This is the type of AI that we create at Aurora. An example might be an AI that recognises faces. It won’t recognise cars, diagnose cancer, or predict the weather. But the next one we create could do any one of those things.
What is Deep Learning?
Deep Learning (DL) is a term that has been given to the breakthrough techniques by which a neural network is trained to perform a specific task. The “Deep” part refers to the number of neuron layers being trained, which may contain billions of interconnected neurons. To provide the vast computing power necessary, Aurora has engineered a specialised Deep Learning platform to accelerate training cycles and shorten development from months to just days.
Do I need specialist hardware?
No, a completed Artificial Intelligence (AI) can run on any standard PC or laptop. To create the AI in the first place is another matter. Aurora has built a high-performance cluster of GPU (Graphics Processing Units) nodes, engineered specifically for Deep Learning.
Your AI will be created using more than 60 CPUs and over 20,000 parallel Deep Learning cores, but the final packaged solution will run on standard hardware and does not even require a single GPU.
How is this different from Machine Learning?
Machine Learning is a field of computer science that includes a range of techniques for analysing data and modelling complex systems to make predictions. Deep Learning is a new, revolutionary technique in the field of Machine Learning that is making waves throughout the industry.
Traditional Machine Learning techniques do not necessarily make use of neural networks, whereas Deep Learning explicitly trains a large (deep) neural network to solve a specific task. This technique requires the innovative combinations of specific concepts to become effective, such as Rectified Linear Units, Dropout, Convolution Neurons, Weight Decay and Stochastic Gradient Descent.
Want to learn more?
Please send you request for information through our contact page.
Alternatively, if you’re attending PTE 2019 in London ExCel (26th – 28th March), you’ll find us on stand #1090.