Best AI and ML Applications of the Week—March 22, 2018
Brazilian Forest Warriors Fight Back with ML
The Tembé people in Brazil work hard to stop invaders from cutting down their trees. Their 30 forest ranger warriors have been patrolling the border of their land to protect it from abuse, but this approach has met conflict—deforesters sometimes bring armed guards who shoot and kill the Tembé forest rangers. With the help of Topher White from Rainforest Connection, this tribe now uses machine learning to identify sounds of logging trucks and chainsaws in use on their property. Then the system sends real-time alerts to tribal leaders, allowing them to immediately contact the authorities. The devices they’ve mounted on various trees listen for sounds as far as a kilometer away, much farther than what a human ear can pick up. Now, the forest rangers can use the copious amounts of time they used to spend patrolling the land toward other productive activities. Watch the video here:
Billions of Free-Floating Sensors
As a kid, you may have wondered how a large Baleen whale could survive on plankton, tiny animals in the ocean. It seems crazy that such small foodstuffs can fill up a whale. But those tiny foodstuffs do much more than feed whales. They also act as ocean, lake, and river sensors because they’re extremely sensitive to water shifts. If their behavior changes, it can indicate changing water quality or changing temperature. But we need to know more about plankton reactions to water changes to understand what these tiny creatures may be telling us. Maybe there’s been an oil spill, a chemical runoff, or a looming red tide.
IBM has an idea. It’s created a small microscope for in-ocean observation of plankton behavior. Until now, scientists generally observe plankton behavior in a lab, an expensive activity, but with IBM deploying their ocean microscopes around the world (or at least that’s the current aim), observing plankton in their natural habitat will help us understand plankton behavior better, hopefully leading to identifying when something goes wrong in our oceans. Watch the video here:
When Your Car Knows More about You than You Think
One company’s new product, an in-car camera that identifies and monitors your every emotion, has some interesting goals for future product iterations: not only is a future iteration aimed at identifying when someone drunk gets into a car, but as cars go autonomous, this product’s creators plan to have it identify your age, race, and gender—as well as predict what your relationship is with others in the car. Eventually, they believe this tech will connect with other smart devices to collectively identify your ideal user experience, personalizing music, adjusting car temperature and lighting, and changing car TV content.