Research papers are coming out too quickly for anyone to read them all, especially in the field of machine learning, which now attacks (and produces documents in) virtually every industry and enterprise. This column aims to gather the most relevant recent findings and journals – particularly in but not limited to artificial intelligence – and explain why they matter.
The topics in this week’s Deep Science column are a real catch bag ranging from planetary science to whale tracking. There’s also some interesting insight from tracking how social media is used and some work trying to move computer vision systems closer to human perception (good luck with that).
The ML model detects arthritis early
One of the most reliable use cases of machine learning is to form a model on a target model, say a particular shape or a radio signal, and release it over a vast body of noisy data to find possible successes. that man could struggle to perceive. This has proven useful in the medical field, where initial indications of serious conditions can be spotted with enough confidence to recommend further testing.
This arthritis detection model looks at X-rays, similar to doctors doing this work. But when it is visible to human perception, the damage is already done. A long-term project that followed thousands of people for seven years made for a great set of training, making the first almost imperceptible signs of osteoarthritis visible to the AI model, which predicts it with 78% accuracy three years out.
The bad news is that knowing early doesn’t mean it can be avoided, that there is no effective treatment. But that knowledge can be applied to other uses – for example, tests much more effectively than potential treatments. “Instead of recruiting 10,000 people and following them for 10 years, we can only enroll 50 people who we know are going to have osteoarthritis … Then we can give them the experimental drug and see if it stops the disease from developing,” he said – author Kenneth Urish. The study appeared in PNAS.
Use an acoustic monitor to prevent whales preventively
It’s amazing to think that ships always encounter big whales and kill them on a regular basis, but it’s true. Voluntary speed reductions haven’t been very helpful, but a smart, multisource system called Whale Safe is being put into play on the Santa Barbara Canal that could give everyone a better idea of where the creatures are in real time.
The system uses an underwater acoustic monitor, near real-time forecast of probable feeding areas, current warnings and a machine learning process (to quickly identify whale calls) to produce a forecast for the presence of whales. along a certain course. Large container vessels can then make small adjustments well ahead of time instead of trying to avoid a last-minute pod.
“Predictive models like this give us a clue to what’s ahead, like a daily weather forecast,” said Briana Abrahms, who led the effort from the University of Washington. “We are using the best and most recent data to understand which habitats whales use in the ocean, and therefore where whales are most likely to be while their habitats change every day.”
Incidentally, Salesforce founder Marc Benioff and his wife Lynne helped establish the UC Santa Barbara center that made it possible.