(ORDO NEWS) — When it comes to making real-time decisions about unfamiliar data, existing AI and machine learning technologies don’t come close to matching human skills.
That’s why NASA scientist John Moisan is developing A-Eye AI.
Moisan, an oceanographer at NASA’s Wallops Flight Center, said his A-Eye AI will use a moveable sensor.
After analyzing the images, the AI will not only be able to find known patterns in new data, but also guide the sensor to observe and detect new features or biological processes.
“A truly intelligent machine must be able to recognize when it encounters something really new and worthy of further observation,” Moisan said.
“Most AI applications are mapping applications trained on familiar data to recognize patterns in new data. How do you teach a machine to recognize what it doesn’t understand, to stop and say, “What was that?
Let’s take a closer look.” This is a discovery.”
According to Goddard expert James McKinnon, finding and identifying new patterns in complex data is still the domain of scientists.
Scientists analyze large datasets by looking at visualizations that can help uncover relationships between different variables in the data.
Moisan intends to primarily focus on interpreting images from Earth‘s coastal regions. He hopes to achieve this by training AI through observations of previous flights over the Delmarva Peninsula.
The built-in AI will scan the collected data in real time to find meaningful features, and then drive the optical sensor to collect more detailed data.
AI could play a big role in the future exploration of our universe.
Sophisticated computers trained to recognize chemical signatures that could indicate life processes or landscape features could add value to scientific data from exploration of the moon or deep space.
McKinnon believes that modern AI is not ready to make critical decisions. “The worst thing for a scientist is to throw away data that could be valuable.
AI can prioritize which data to send first, or have an algorithm that can draw attention to anomalies, but in the end it is the scientist who studies that data that will lead to discoveries,” he said.
Contact us: [email protected]