Innovations in data and AI at NTT’s R&D Forum | NTT DATA

Wed, 19 February 2020

Innovations in data and AI at NTT’s R&D Forum

Data is laying down the foundation of a new, smarter world.

by Tom Winstanley, Vice President, New Ventures & Innovation, NTT DATA UK

Data is laying down the foundation of a new, smarter world.

It’s at the heart of the digital transformation of enterprises and public services. Businesses have woken up to the competitive advantage that better use of data provides – from reducing costs to obtaining better insights into their customers.

RexSense standHowever, using data at scale requires suitable infrastructure, and that’s a struggle for most organisations.

Just having the data isn’t even close to being enough. As data volumes increase, we need to create smart infrastructure to make the best use of it. This means taking advantage of technologies like IoT, edge computing and efficient cloud data management.

We increasingly need to employ AI and automation technologies to make sense of the data – and the simple truth is most organisations aren’t there yet.

Data and intelligence

At the R&D Forum in November, we showcased our answers to the two core challenges of modern enterprise technology: data and intelligence. When they work together in harmony, the results are remarkable.

In one connected car demonstration, NTT DATA showcased its RexSense and other cutting-edge speech processing technology for improving voice interactions with vehicle systems. It combines an intelligent microphone, which reduces road noise and car music, with software that distinguishes between questions, commands and statements.

RexSense understands the difference between a command to turn up the air conditioning and a request for local restaurant recommendations – and can relay questions it can’t answer to a remote cloud server for detailed answers.

Examples like this – where different technologies from across NTT Group are combined into a single solution – were regular occurrences at the R&D Forum. By leveraging the diverse research capabilities available within our organisation, NTT DATA is able to draw on the latest and most advanced research in the world and deliver it to our clients.

Here are two NTT DATA technologies that were showcased at the R&D forum – which demonstrate the combination of data and intelligence to create new and innovative solutions for our clients:

1. Explainable AI

When NTT DATA introduced its AI ethics guidelines last year, one of the core principles is that we develop fair, reliable, explainable AI. This means that the reasoning behind machine-derived decisions can be presented to humans. By opening up the decision-making process, it helps to counter against unintentional biases that can build up in AI.

Explainable AI also makes the decision more valuable in certain scenarios – for example in medical diagnoses. NTT DATA has been developing a Smart Alert Solution for Intensive Care Units in hospitals that monitors vital signs and gives doctors warnings for specific complications such as septic shocks, where early diagnosis is critical.

The explainable AI portion of the technology describes the combination of factors that makes the machine believe the patient is at possible risk, meaning the doctors can treat the patient appropriately.

At last year’s R&D Forum and NTT DATA Sessions, another explainable AI solution was demonstrated for analysing packages, a technology developed for applications such as the import inspections. This is a task that involves the imaging of closed parcels and bags to ensure safety and compliance.

NTT DATA enhances the traditional X-ray inspection by using AI-based object recognition. This explainable AI solution labels objects like bottles, cameras, watches and knives. The information is displayed to inspectors and they are alerted to potential areas of concern – improving accuracy and reducing fatigue.

2. Automatic Correction for Machine Learning

For any machine-learning derived algorithm to develop high levels of accuracy, it must be able to update itself and its judgement logic during operation. However, at present, once an algorithm has been trained, the accuracy level often declines out in the field.

AICYCLE is an NTT DATA technology that improves prediction accuracy by continually and automatically re-evaluating and updating the algorithms – preventing the degradation in predictions.

It’s a technology that’s already proved its value in the world of manufacturing. When the NTT DATA team recently assigned AICYCLE the job of detecting defects and process improvements within the aircraft engine turbine manufacturing process at Mitsubishi Heavy Industries Aero Engines, it was effective in reducing the downtime rate by 47%. It also managed to achieve a 25% reduction in manufacturing line stoppage.

Entering a new era

The Economist calls it the “Age of AI”, the World Economic Forum describes it as the “Fourth Industrial Revolution” – this new era is about bringing together data and intelligence.

We’re working at the cutting edge of this technology, ensuring the accuracy of algorithms, and in turn, data’s reliability and meaningfulness in the real world.

AI is a huge growth area for our work in consultancy and systems integration. NTT DATA is using its world-leading research and global innovation to help our clients transition to a smarter world.

How can we help you

Get in touch