The latest weapon in the fight against illegal fishing? Artificial intelligence

Date: November 20, 2016

Source: The Guardian
Author: Mary Catherine O’Connor

Facial recognition software is most commonly known as a tool to help police identify a suspected criminal by using machine learning algorithms to analyze his or her face against a database of thousands or millions of other faces. The larger the database, with a greater variety of facial features, the smarter and more successful the software becomes – effectively learning from its mistakes to improve its accuracy.

The latest effort to use artificial intelligence to fight illegal fishing is coming from Virginia-based The Nature Conservancy (TNC), which launched a contest on Kaggle – a crowdsourcing site based in San Francisco that uses competitions to advance data science –earlier this week. TNC hopes the winning team will write software to identify specific species of fish. The program will run on cameras, called electronic monitors, which are installed on fishing boats and used for documenting the catch. The software will put a marker at each point in the video when a protected fish is hauled in. Inspectors, who currently spend up to six hours manually reviewing a single 10-hour fishing day, will then be able to go directly to those moments and check a fishing crew’s subsequent actions to determine whether they handled the bycatch legally – by making best efforts to return it to the sea unharmed.

TNC expects this approach could cut review time by up to 40% and increase the monitoring on a boat. Despite rules that call for government-approved auditors to be stationed on 5% of commercial fishing boats in the Western and Central Pacific, in practice the auditors are found only around 2% of the fishing boats, including tuna long liners. As a result, fishermen sometimes keep protected fish that they hook – including sharks that are killed for their lucrative fins.

In the Pacific’s $7bn tuna fishery, illegal, unreported and unregulated (IUU) fishing not only harms fragile fish stocks, it takes an economic toll of up to $1.5bn. The impact shows up many ways, including lost income for fishermen in the legal marketplace and harm to the tourist economy that sells snorkelers and divers the opportunity to witness protected species in the wild.

Worldwide, cost estimates related to IUU reach $23bn annually, and the take represents up to 20% of all seafood. Using technology to track and prevent illegal fishing presents an opportunity for technology companies as the fishing industry seeks ways to comply with the growing demand for transparency from governments and consumers.

“If using facial recognition software to track fish were easy, we’d already be using it,” says Matthew Merrifield, TNC’s chief technology officer. Whereas images from security cameras installed inside banks or other buildings are consistent and predictable, “the data from (electronic monitoring) cameras on boats is dirty, because the ships are always moving and the light keeps changing”.

Because of the “dirty” data, it will not be easy to write a facial recognition software that can accurately spot protected species when the variable conditions on the high seas could lead to blurry images on the video.

Given those challenges, it’s too early to know how large this market will grow, or how quickly. While the use of artificial intelligence to reduce illegal catch is relatively new, the Kaggle contest isn’t the first time it is being applied to the fishing industry.

San Francisco-based startup Pelagic Data Systems (PDS) has developed technology that illuminates the activity of some of the 4.6m small-scale commercial fishing boats that ply coastal waters around the world. Using data from a UN’s Food and Agriculture Organization report, PDS estimates that roughly 95% of those boats don’t have the types of communications and tracking radios that larger boats are required to have, partly because the boats are too small or lack the power source to run the radios.

PDS installs a solar powered radio with an integrated GPS receiver and cellular modem on boats. The company collects the location data and analyzes it to create a map to show where the boat traveled and deduce its activities, such as where it stopped to set out nets or other gear and where and for how long it hauled in a catch. This data is vital because it shows whether the boat fished inside or outside marine protected areas. The device doesn’t have an on/off switch, a design to prevent a fishing crew from tampering with data collection.

The software also generates heat maps to indicate where the heaviest fishing activities are taking place within a coastal region. By pairing that data with the movements of the boats, PDS can also estimate the quantity and even the size of the fish pulled from those waters, says Dave Solomon, CEO of PDS.

The company sells its technology to governments, nonprofits, academic researchers and companies in the fishing industry, and expects the number of boats installed with its device to reach 1,000 in regions such as West Africa, North America and Mexico by the end of the year, Solomon says. Some of his customers install the devices in the boats of their suppliers for another reason: to win over customers by demonstrating transparency in fishing practices.

Another effort to use data to fight illegal fishing comes from the nonprofit SkyTruth, which tracks the movement of large ships by mining data broadcast by ships and collected by satellites. Its technology is used by Global Fishing Watch, which is backed by Google, Oceana and the Leonardo DiCaprio Foundation. SkyTruth’s data helped the island nation Kirbati to bust illegal fishing operations.

But Kaggle has a habit of taking on unusual technical challenges. Earlier this year, it launched a contest with State Farm to develop machine learning software, to be embedded in dashboard cameras, to classify a driver’s behavior, such as being distracted by a smartphone when behind the wheel.

Kaggle, with a membership of 650,000 data scientists, hasn’t tackled an environmental problem before. But its CEO, Anthony Goldbloom, thinks the TNC contest could represent the start of environmental competitions on its site because scientists from government agencies and academic institutions are collecting a growing amount of field data using cameras and sensors.

TNC contest attracted 44 teams within the first day. Each team has five months to submit its software.

While the contest presents an appealing opportunity to do something good for the environment, it doesn’t promise a big payoff. That will make it difficult for software developers and data scientists to raise venture capital to fund their efforts.

“Silicon Valley only invests in places with big money [potential],” says Andrew Bosworth, vice president of ads and business platform for Facebook and a board member of land conservation group Peninsula Open Space Trust. “Plus, everyone underestimates [environmental] challenges. Going to the moon is easier than tracking fishing. It really is. So these are big challenges without financial incentives to solve them.”

But, he adds, Silicon Valley does provide important undergirding for using technology to solve environmental problems. Bosworth argues that the advancement in core technologies behind things like multiplayer gaming software and smartphone apps has propelled the rise of machine learning and artificial intelligence and lowered the development costs over time.

The winning team of the contest will earn a prize of $150,000. Then, as part of its campaign to reduce bycatch and illegal fishing in the region, TNC will work with the governments of Palau, Federated States of Micronesia, Solomon Islands and Marshall Islands to install the software, for free, on the electronic monitors of selected fishing boats.

If the software proves effective in reducing the labor costs and improving the accuracy of identifying protected species, then it could become a standard feature in electronic monitors. TNC will own the intellectual property of the winning software and make it free to the equipment makers, which include Satlink and Archipelago. The software could become even more widely used if large retailers such as Walmart begin to require electronic monitors on their vendor’s fleets.

But it is still early days for policing the fishing industry. For Melissa Garren, chief scientific officer of PDS, that means the market potential is huge. “We should be treating the oceans more like we treat airspace,” she says. “If we had this lack of visibility in the skies, it would be nuts.”

Posted on Categories GlobalTags