Introducing: Nearmap for Post-catastrophe Response
Post-catastrophe imagery and AI-derived property damage and condition data unite to help insurers process customer claims more efficiently.
Post-catastrophe imagery and AI-derived property damage and condition data unite to help insurers process customer claims more efficiently.
The appeal of Artificial Intelligence for geospatial planning is now a reality with the recent launch of Nearmap AI. Dr. Mike Bewley, Senior Director of Artificial Intelligence systems at Nearmap, is here to provide greater detail on how you can bring these valuable, pre-programmed insights into your workflows.
Before we jump into our conversation with Dr. Mike, don’t forget that you have a chance to hear from him at our annual user conference, Nearmap NAVIG8 2020. You’ll get an opportunity to qualify for an exclusive hands-on experience of Nearmap AI!
Machine learning and artificial intelligence have been an interest for Dr. Mike since his student days at Sydney University. In 2007, while working on an undergraduate project detecting Alzheimer's from 3D brain scans, Mike first experienced what he describes as "the wonder" of machine learning. It not only sparked Mike's passion for artificial intelligence, it won him the university medal for his undergraduate thesis.
While working on his PhD in marine robotics, Mike applied machine learning to support a project around finding better ways to study underwater ecosystems. Traditionally, researchers manually counted and catalogued marine life within rope grids laid on the seafloor, extrapolating their observations to 'guesstimate' findings for broader areas of the environment. Foreshadowing his future work with Nearmap, Mike joined a team that had introduced an autonomous underwater vehicle fitted with cameras to capture vast areas of the seafloor in hundreds of thousands of images. He worked on an AI system to automatically 'label' content in the images according to a defined scientific taxonomy, from which meaningful and accurate data could then be extracted.
Fast forward to now and Mike remains as excited by AI as ever. He says it can be easy to be distracted by the "razzle-dazzle" of AI and the challenge for those working to apply it is three fold. "First, we need to look at harnessing the sheer volume of data we can extract from our high-resolution aerial imagery. Next, we need to understand how that data can meet customers' needs and how to provide that data in forms that they can use. Above all is quality of the data."
With the theme and promise of Nearmap NAVIG8 2020 being 'Truth is in the Detail’, you’ll want to tune in to understand how and why AI is delivering on that promise. "We have algorithms looking at algorithms and people looking at algorithms to refine the system that creates reliable and powerful models and data sets."
It was only a few years ago that Nearmap users began asking how we could help them to automate repetitive tasks in analyzing aerial map images or to help find those figurative needles in haystacks without painstaking manual examination and analysis. It was from these requests that Nearmap AI was born.
"I've often said that it's our customers who are best at disrupting themselves," shared Mike. "It's our job to show them what's possible, then they can take the data and find new ways to turn AI data into their own insights and business advantage by changing the way they do things."
Mike notes that the value that can be extracted from Nearmap AI data can be different for every user. Consider, for example, one of Nearmap AI's popular data packs which essentially identifies the footprint of buildings within an area captured by aerial photography. While one user may simply want to identify the number of buildings, another may want to understand the liveable area of dwellings. The same data may be useful for developers to identify undeveloped land lots or for insurance businesses to assess replacement costs. Urban planners can compare data sets to understand change over time and accurately inform their forecasts and predict the nature and scope of future redevelopment.
While Mike insists that it is customers that will drive change, it's his dedicated AI team that will deliver it. Part of his presentation at Nearmap NAVIG8 will take us behind the scenes to meet the talented AI team.
Currently numbering 15, Mike says his team covers a broad spectrum of talent and specialty. "We have data scientists who are committed to making data meaningful for our customers and at the other end we have machine learning engineers who work at how we can process data at ever bigger scale and ever faster speed. We have academics and seasoned industry specialists, PhD graduates from places like Oxford and Cambridge and current science students and graduates bringing in new ideas."
While Nearmap AI has already changed the way customers are working around the world, Mike insists that "the team is just getting started and we have at least ten years of ideas."
