Tuesday, August 14

Alt-data from outer space: how AI and satellites are changing the options

LinkedIn Google+ Pinterest Tumblr +

Commercial efforts in space exploration have reached a tipping point, making the industry cheaper and satellites smaller for remote sensing capabilities, said presenters at the annual Ilan Ramon space conference in Tel Aviv.

Punctuated by Elon Musk’s SpaceX, “the new space” is all about “small” and “business models” according to one speaker describing the changes taking place.

Major General Robert Skinner, deputy commander of US Air Force Space Command, said that the commercial industry is out in front and building faster than most governments can keep up.

“The slope of the curve for progress in the space industry seems to be changing from steep to vertical,” he said.

The stats are eye-catching: 2017 set a record of almost USD$4 billion in non-government equity investment in commercial space companies. In 2013, the world saw a shift in the launch business as the number of satellites almost doubled. This has now become average, said Skinner.

What’s in it for finance

Another meaning of the “new space” is opening space and satellites to new audiences, said Ori Zeisel (via a translator), director for business development and marketing at ImageSat International, a provider of high-resolution satellite earth imager and data analysis to better understand and act on geopolitical, environmental, and economic realities.

ImageSat presented case studies of “Intelligence-as-a-Service”, which included using algorithms and satellite imagery to track terrorist activity in West Africa and illegal fishing. Hedge fund clients for IaaS are mostly in the oil and gas space, said a spokesman of the company speaking on the sidelines of the conference.

ImageSat International’s Ori Zeisel presenting at 13th annual Ilan Ramon Space Conference

Angieszka Lukaszczyk is the EU policy director at Planet, a firm which started out in a garage and now designs, builds, and launches satellites faster than any company or government in history by using lean, low-cost electronics, and design iteration.

She noted that finance and Insurance companies are one area where satellite data can be and is being used.

Planet’s first mission was to image the whole world every day making change visible, accessible, and actionable. For those curious what that looks like, check out their API.

The company’s next mission is to index earth in the same way that Google indexed the internet so that people can query what’s happening on the planet over time.

Another relatable concept came from Ramon Chips, which makes processors, computers, systems, and software for space. The company is trying to create apps on satellites to operate just like those on smartphones.

Can we predict the future yet?

Predictive analytics is one of the major points of interest in finance, and making sense of the images streaming in from the universe isn’t your mother’s linear regression.

Ami Daniel, co-founder and CEO of Windward, said that data platforms are everywhere but analysts are confused.

The solution, he said, is “vertical AI”, which means combining domain expertise with data science.

“You will see companies who focus on specific verticals and build vertical specific algorithms to bridge that gap,” he said.

Windward’s vertical AI is maritime data analytics, but the idea could be extended to energy and transportation, for examples.

One of the most known use cases of satellite data is counting cars outside of Walmart, which is used as a kind of benchmark for revenues of that specific location. Adding up all the branches, you get the revenue, right?

Well, no.

“We’re seeing that counting is not enough, measurement is not enough, we need to take the next step towards prediction: risk, relationships, what’s going to happen,” he said.

“What we see is that combining domain expertise and data science, and using graph theory to model how vessels behave, we can unravel relationships between vessels.”

Using AI and domain expertise, Windward is predicting next year’s accidents for vessels, like Lloyd’s of London insurance firm predicts next year’s oil spills, explained Daniel.

And then there are global supply chains: what does bad weather in India do for Japan trade? It’s a problem that insurance firms are only too keen to solve.

“If we model the world as a network of entities interconnecting and interfacing every day, we could see what will happen tomorrow, next week, next month and expand our horizon,” he said.

Climate change and environmental events (floods, earthquakes, etc) are within the purview of risk and asset management, said Massimo Comparini, CEO, e-GEOS, an image intelligence firm.

Big data and social data will define the severity of such events, and algorithms will be tools with Platform- and Software- as-Services models, he explained.

Quantum leap

Eugene Tu, director of NASA Ames research centre, provided updates on quantum computing in space exploration, and for once, it seemed like finance might actually be ahead of the curve. Of course, no financial player is facing quite the same problems as NASA.

Quantum computing is worth investing in the early stages because of the promise that it can handle large optimization problems, explained Tu.

Classical computing gets stuck in low optimization, whereas with quantum computing, the promise is you can really truly find the global optimization, he explained.

NASA is interested in quantum computing for large problems; data analysis and data fusion; air traffic management; mission planning, particularly in areas of unknown environments; and anomaly detection and decision-making.

The US space agency is working with D-Wave’s quantum annealer as a research device, and estimates some five years to see what kind of breakthroughs may happen.


Read MarketBrains’ interview with D-Wave’s CEO Vern Brownell


But just like supercomputers have never replaced wind tunnels at testing facilities, quantum computing will not replace supercomputing.

“It may take a supercomputer to help analyze the data that comes from a quantum computer, so the two will work hand in hand,” he said. “Another way to think about it is a quantum computer is a node off of the supercomputer specifically designed to solve problems that are around optimization.”

The conference also covered a number of other issues such as advancing social progress and cybersecurity.

Luc St. Pierre, chief of the space applications section of the United States Office of Outer Space Affairs (UNOOSA), spoke about an alignment of commercial efforts with the UN Millennium Development Goals, and the tremendous need for developing countries to access data for crisis management in the wake of natural disasters, for example.

Planet’s Lukaszczyk warned against the excesses of “techno-utopianism”, but business models, she said, should promote social good.


About Author

Leave A Reply