On the shoulders of giants

Winning ways with artificial intelligence

A Capgemini team in Norway has developed an AI solution in support of a vital conservation and environmental sustainability mission. Team members Hartvig Johannson, Pedram Sherafat, and Simen Norrheim Larsen discuss the essentials behind their approach to innovation

Behind great scientific innovations are teams of experts whose names don’t always make headline news. Yet their skills help forge the transformations that come to benefit the world. Their job is to build on the work of others, harness the power of all the knowledge and tools at their disposal, and apply it to solve a specific problem.

In this way, data scientists at Capgemini are using technological innovations to understand climate change better and to help create a sustainable future for life on Earth. One initiative that brings them together to look at how they can help create a better future is the Capgemini Global Data Science Challenge.

The 2021 challenge is now underway, with rival teams hoping to match the innovations made in 2020. The team that won last year’s competition developed a cutting-edge data solution that is set to play a part in protecting the future of an endangered species – sperm whales. Their work will help biologists map whale migration patterns and monitor any changes in the species’ behavior that might have arisen from environmental factors, such as climate change and warming ocean temperatures.

For the 2020 challenge, the Norway-based Capgemini team used Amazon Web Services (AWS), artificial intelligence (AI), and machine-learning technology to develop a tool that can accurately identify sperm whales via photo image processing. In turn, this will help scientists track the whales’ migration routes, look at the social structure of the sperm whale groups, and protect the whales’ natural habitats. It’s an innovative use of AI that will help to boost species conservation and environmental sustainability.

How did they achieve it? According to team members Hartvig Johannson, Pedram Sherafat, and Simen Norrheim Larsen, creating a winning project that can deliver real-world impact is about applying tried-and-tested principles of innovation.

Principle 1: Use your experience

“For a start, when we came together as a team we all had experience in the field of AI and analytics as part of the same insights and data department at Capgemini,” says team member Hartvig. “Challenges like this are great opportunities to learn and apply that experience to practical, real-world use-cases.”

“I’m also personally interested in nature and the environment, so it was exciting to see how we could take what we know from computer science and use it in a completely different area.”

Principle 2: Start with the data

While AI, machine learning, and AWS might supply the buzzwords for a tech innovation story, the foundations for the team’s success were built on understanding the available data.

It begins with a scoping exercise.

“With what we do, the best approach to solving a problem is always to start with the data,” says Pedram, another member of the Capgemini team.

“In-depth data analysis must come before applying machine learning or AI,” adds Hartvig. “You have to really understand the problem and what you are trying to solve, as well as the potential the data provides, and also its limits.”

Principle 3: Keep it simple

According to the team, tools and technology can sometimes be seen as a fix-all. But that is rarely the case. Instead, it is better to tightly define what you want the tech to achieve.

“You can’t simply sprinkle some AI on top of a problem to expect a good outcome,” says Simen, the third team member. “The misconception is often that you can, but any good modelling is built on the foundations of knowing the domain and what the technology is capable of.”

Pedram adds: “It helps to have a specific challenge. With ours, it was to identify the individual ‘fingerprints’ of sperm whales through images of their flukes, or tail fins. A defined challenge like this is where machine learning really shines.”

Principle 4: Build on the work of others

Finally, the team agrees that innovations often build on the work of others.

“People have been identifying whales manually for years,” says Hartvig. “A lot of ground was already prepared, so we could come in and add a technical solution to build on what had previously been achieved. It has been a great way to make the most of machine learning and AI.”

The scientist Isaac Newton famously said, “If I have seen further it is by standing on the shoulders of giants.”

It seems that principle still holds true in today’s world of AI and machine learning – and even when the innovation in question is about safeguarding the future of the giants of the sea. When passionate people and smart technology combine in clever ways, it can help Capgemini clients succeed in their own specific sustainability challenges – and deliver a more sustainable future for the planet as a whole.

Discover more about the project here or watch a short video on Reuters’ Twitter feed.

Using AI to protect whales

Working with Lisa Steiner, Capgemini and AWS develop a new, self-learning AI solution that can more effectively identify matches between images of sperm whale tails and make it easier to draw a picture of migration going back decades

COP26: making informed decisions

26th United Nations Climate Change Conference of the Parties, COP26, marks the deadline for governments to set out their more ambitious climate change goals following the Paris Agreement signed at COP21 six years ago.