Deep learning gives us more intelligent apps
Whilst artificial intelligence (AI) is considered an unavoidable evolution to help businesses in their digital transformation journey, the technological revolution – also known as “cognitive information” – is making its move on the market. This will offer companies, IT Managers, and specialist professions, new and more “intelligent” services, particularly through the use of machine learning and deep learning.
Of course, the use of these apps is only in the early stages and services are still emerging. But image recognition, facial or even textual recognition are all examples of AI technologies drawing on machine and deep learning, with applications across various industries such as security, defence, banking, and healthcare. These technologies, sometimes linked to the Cloud and enabling flexible usage and exploitation, are generating evermore intelligent services. Employees will be able to take advantage of these services, even if they will have to learn to adapt. IT Managers will also find new resources for the company’s internal users within these new intelligent services.
Machine learning and deep learning: lab-born solutions
Machine learning may be defined as automatic learning that enables the computer to learn from its own experiences. Deep learning is a type of automatic learning application, and, more specifically, an evolution of systems known as “neurone networks”. Deep learning can solve more complex problems using a high number of neurones and networks that come together to form a system, which is designed to function in the same way as the human brain.
These are lab-born solutions. All the biggest players in the IT and digital world – including Microsoft, of course – are working on such solutions. We might take this opportunity to draw attention to Azure’s solution, which, thanks to Cloud computing, makes it possible to integrate these technologies into different applications. Similar to with Sopra Steria, the Azure ML module enables the development of applications with integrated machine learning, such as an innovative pharmacy management solution for example. As for Microsoft Cortana, the PA solution, this can also be used directly within Azure.
API systems bring solutions to business
We are seeing this domain forward. These technologies generally make it easier to address the issues with recognition (image, textual, facial, etc.), to roll them out and get them underway. Through this we are seeing the birth and rise of numerous new applications, with the most well known being voice recognition available on Smartphones. Voice and facial recognition have taken off, and man-machine interaction has been now greatly improved. We are also seeing the rise of fingerprint recognition, used more and more across different industry sectors. Everything is loud and clear for the user.
All that’s left for us to do is implement these techniques. How, then, can we integrate these different technological layers? Well, Microsoft has integrated AI modules directly into Azure, namely machine and deep learning that developers are able to exploit quite easily. Microsoft has also made its deep learning library available as an open source, and created its own deep learning engine, based on Minecraft.
You get the picture. These technologies are becoming much more available, open, and perfectly functional. Moreover, they are newly available via API’s, such as the Microsoft Cognitive Service, and ready direct for integration by developers looking to integrate recognition services such as images, faces, smiles, voice intonation, and even emotions.
An impact on professions
This impact is, of course, part of the general underlying trend of digital transformation. Amongst the drivers behind DT and alongside mobility, IoT, and smart machines, data science (another name for AI) finds its rightful place. Its integration is quite simple and the real issues that we may encounter concern the machines themselves and their capacity to learn. But these problems are disappearing naturally over time; such is the principle of machine learning.
The main hold-ups remain human, notably in the reluctance to accept the systems, the impact on professions, professions that are changing, and job titles which have to be redefined.
Machines may be learning, but the human element still has a long way to go…