Artificial Intelligence, Machine Learning and Deep Learning
Why should you use the Artificial Intelligence?
AI is becoming an integral part of our society. This article presents how it can be applied, whether in the B2C or the B2B context, and the relationship between AI, Machine Learning and Deep Learning.
Artificial, Machine Learning and Deep Learning are becoming common words: these technologies are now part of our everyday live. Not only for those who work in the business arena or for individuals on the cutting edge of technology, but truly for society at large.
Today smart algorithms are able to teach themselves, to recommend what purchases to make, what music to listen to, and what route to take. They pull data from images and documents and they even interact with us through smart speakers that recognize our voice. These are just a few examples of how Artificial Intelligence is now a part of our daily lives.
A very interesting tension has emerged from a study conducted by IBM*, Roadblock to Scale: The Global Sprint Towards AI: Artificial Intelligence may well represent the greatest economic opportunity around us – according to PWC**, is set to add 16 trillion dollars to the global GDP by 2030 – but its implementation of remains fairly limited.
Furthermore, although AI is already improving productivity for a host of companies, thanks to its ability to automate repetitive or menial tasks – with an attendant reduction in error rates – there are still high barriers to entry with this technology, including lack of expertise, equipment, and trust.
Artificial Intelligence, Machine Learning and Deep Learning: what is the difference?
Artificial Intelligence can be considered a real technological revolution, being able to change both commerce and industry.
What is the difference among Artificial Intelligence, Machine Learning and Deep Learning?
The Artificial Intelligence is a system that simulate the function of the human brain, able to make decisions based on the analysis of processed data. It makes possible to have machines tackle tasks like voice and face recognition, as well as to carry out decision-making and predictive processes.
Machine Learning is a particular type of AI and Deep Learning is a subset of Machine Learning. Put simply, they are a set of algorithms and programming methods.
“Machine Learning” is a system able to learn independently and recognize its own mistakes, based on algorithms that analyze data. By learning from its own mistakes, it can make decisions and provide forecasts. The learning process takes place through Big-Data analysis.
Deep Learning o Neural Networks
Deep Learning, just like Machine Learning, is a system capable of teaching itself, and learning from its own mistakes. In addition, Deep Learning use a complex system of neural networks which simulate the cellular behavior of our brain.
Artificial Intelligence as a field of study: designing innovative applications
The Competence Center AI at Intesa (IBM Group) is an innovation hub where this technology is studied, focusing on its integration into existing systems and synergy with other entities, which allows Intesa to design, test, and reproduce business-ready and user-centric applications. Our partnership with the Links Foundation, launched by the Turin Polytechnic and the Compagnia di San Paolo, is a perfect example, indicative of our willingness to take an iterative approach to Open Innovation.
From this collaboration one of the most interesting technological innovations in the field of exponential technologies has born: the development of an original neural networks applied to services offered by the company.
These systems have proven to be extremely useful at the ID-verification stage, as well as with identification or identity-recognition. Smart algorithms are, indeed, able to:
- Automatically match the identity document to the document type
- Improve image quality, using computer-vision logic
- Compare the image content with the video frame, and determine whether the person who made the video matches the person appearing in the ID
- Verify whether the person speaking is the one depicted in the video by analyzing their lip movement.
Where do we find Artificial Intelligence? Examples of use
Artificial Intelligence has a wide array of applications, depending on the need to be met.
On a B2B level, one can exploit this potential by applying its logic to improve the management of processes up and down the Supply Chain. By creating connections among shared documents and data (such as through EDI), one can conduct predictive analysis in a way that streamlines the entire Supply Chain (from managing leftover materials, to forecasting weather and economic risks, to setting new delivery routes).
AI can provide tremendous support in the B2C arena, too. Imagine, for example, onboarding procedures, and the type of identification protocols described above, where AI can recognize the document type and the user’s actual identity, making sure that the person asserting the identity is the one appearing in the ID itself.
An intelligent system able to automatically match the data with AI and Machine Learning; it can also prevent fraud and effortlessly uncover criminal behavior, making transactions more secure and improving customer experience.
For a company that has made innovation and technology hallmarks of its business, one of the most important missions is staying on the cutting edge, continually striving for better performance, integrating technology and human expertise through a sustainable approach to innovation.
* Source: From Roadblock to Scale: The Global Sprint Towards AI, IBM Corporation 2020
** Source: https://www.pwc.com