Artificial intelligence (AI) and machine learning are generating interest from IT leaders eager to discover how the technologies can help their business meet the constant demands of driving down costs, achieving growth, deriving insights from data, making decisions faster and improving efficiency.
But there is a huge amount of hype in this emerging area, and IT chiefs need to separate fact from fiction surrounding AI. It’s vital to understand where AI is already working in real-life situations and focus on what it can do for your organisation.
"AI is not black magic," says Walter Riviera, an AI technical solution specialist at Intel. "Machine learning in banking predictions, for example, has been used for many years, but it is only now that we can unlock data and apply the processing power. AI is allowing us to achieve more complex goals in a doable way."
AI provides new tools to tackle old problems by giving organisations the ability to extract more relevant information from data. IT leaders are entering a period in which many sectors, including banking, security and healthcare, are capitalising on this capability, with valuable innovation lessons to be learned.
A startup in Italy, for example, is using AI-enabled drones to transfer organs between hospitals, making the process more time and cost efficient. AI and deep learning are also being used in airport cameras to improve security and in microphones to improve language translations.
Riviera says there is huge potential in many areas but no such thing as an off-the-shelf solution.
"Start from a business need. What problem are you trying to solve? What data do you have available that could address a business need?" he says. "Start targeting what you need and build on top creatively. The first step is theory, the second is data, and the third is compute and resources."
Once a business has identified the challenge it wants to address, Intel can help.
"We can see what an organisation is trying to achieve, and there are lots of possible scenarios. We can come in with a solution that has worked before, or we can build a new one," says Riviera.
One solution cannot fit all needs, and that is where Intel’s wide portfolio can help. It offers tailor-made solutions from the data centre to the edge, and from low-level hardware to user-friendly software such as OpenVINO and BigDL, Intel’s end-to-end analytics and AI platform.
"Our point of strength is that we come from a generation where people have their own data centre. We can unlock deep learning technology by using the current infrastructure with BigDL on top and preserve the legacy," says Riviera.
From the data centre to the edge
With the new generation of Intel Xeon scalable processors, organisations can now harness compute power and create agile machines for deep learning in the data centre. This evolution can be realised not just in the data centre but also at the network edge.
For example, deep learning is being used for dynamic targeted advertising on digital billboards along the Tokyo expressway in a collaboration between Intel, Cloudian and Dentsu, whereby the system displays different adverts according to the types of car. AI is also being harnessed by Formula One racing to exploit data and make decisions at the edge.
"Whether you decide to enter at the data centre or at the edge, Intel has a solution to help any company out," says Riviera.
"As long as you have your own data, you can start the journey today. We provide software developers with tools to build an AI algorithm. It is potentially not costly at all."
Advancements in AI are helping businesses innovate and create new jobs, he adds. "AI is inventing new ways of doing things. It is an enabling technology and enhancing what we already have and do."
This is particularly apparent in the healthcare sector, where AI is improving diagnostics, discovering cures and improving efficiency. Intel is working on several projects including early identification of osteoporosis and accelerating genome sequencing in research and clinical oncology.
"However much information you collect, you will always need a doctor," Riviera notes. "AI is just providing a new point of view to the decision-maker, who is a doctor. AI provides the tool; it does not replace the human."
Not reinventing the wheel
The security sector can benefit from AI as well. Enormous amounts of data render it impossible for humans to sift through the information to make accurate decisions in areas such as fraud detection and terrorism prevention. In contrast, AI can analyse large volumes of data much faster and with greater accuracy than traditional methods.
For example, Riviera highlights how Intel deep learning technology is being used by China UnionPay to refine fraud detection by replacing its old rule-based model to capture exceptional or uncommon transactions, such as "late-night" transactions (depending on what "late night" might be for different age groups).
"These concepts can’t be captured with a rule-based system, but deep learning can capture these concepts and is more flexible and powerful. It is not reinventing the wheel but heightening security and reducing calls to the bank," says Riviera.
In the rail industry, Intel’s AI technologies are helping a European rail operator predict train delays and improve capacity alongside Resonate’s Luminate digital traffic management system.
The operator manages 50 million customer journeys and 2,000 services per day on 1,300 kilometres of track, with 82 stations. Key benefits include a 9% increase in on-time arrivals, with 92% of trains now arriving on time; increased automation, with an additional 6,000 short-term schedules; and improved station planning.
Across sectors, AI is revolutionising business.
Intel’s message for organisations investigating how AI can help is clear. They can start their AI journey today, and BigDL can be used with existing infrastructure to make it more powerful and up to date.
"No matter what you are trying to achieve, we have got the tools for you, software or hardware," says Riviera. "One solution doesn’t fit all, and we can customise."