The AI ​​helps detect wildlife health problems in real time


During the Spring Festival, Seabirds on the California coast died of domoic acid poisoning, which is caused by the proliferation of harmful algae. An early clue indicates when and where the problem began to spread: rescued California brown pelicans, red-throated loons, and other species began to appear in wildlife rehabilitation centers with signs of neurological disease. However, although they are all over the state map, the interconnection between these centers is not enough to nip the problem in the bud. When a center worker diagnoses a sick bird, other people along the other 40 miles may not know the information.

Therefore, researchers at the University of California, Davis recently tested an early detection system that uses artificial intelligence to classify admissions to rehabilitation centers, hoping to inform wildlife agencies and researchers about marine birds and many other species of animals. Warning of serious problems. Their system scans intake reports generated by 30 centers in California, listing the animal’s species, age, reason for admission, and diagnosis. The artificial intelligence then uses natural language processing to classify the reports, looking for patterns in hospital admissions related to certain diseases and injuries.

The researchers used five years of data and more than 200,000 records to establish a baseline of how often these conditions usually occur. When the system detects an anomaly-an unusually high number of cases in a given species-it will automatically create an alert and send it to wildlife experts via the system dashboard, email or text message. Since the system can process the admission data of the rehabilitation center in just one or two days, it can generate a “pre-diagnosis” alert, which is faster than waiting for the diagnosis to be confirmed.

In July, the team Published a paper describing their system test In the magazine Proceedings of the Royal Society“We hope to use the data in an aggregated form to better help the recovered people see a bigger picture, rather than what they see at their respective centers,” said the chairman of the Wild Neighbors Database Project and one of the authors of the study. Said Devin Dombrowski. Paper.

In a one-year pilot study, the system identified several models that showed major problems. A large number of seabirds with neurological symptoms such as head twitches and body tremors triggered the alarm. After autopsy, it was found that these birds, including the western thrush, a black and white waterfowl, were found to be suffering from domoic acid poisoning. A few months ago, the high rate of admission to clinics in the San Francisco Bay Area due to rock pigeons with neurological disease symptoms triggered another alarm.Further investigation to determine the parasite Sarcocystis As the reason.

Study co-author Terra Kelly, a veterinarian and epidemiologist at the University of California, Davis, combined the system with Symptom monitoring For people, it uses electronic health records to monitor public health issues, such as influenza outbreaks, opioid overdose, and the spread of Zika virus. Coronavirus diseaseShe pointed out that animal alert systems can also benefit human health. She said, “Wild animals can be used as early indicators of diseases such as West Nile virus.”Diseases that have been killed More than 2000 people According to the Centers for Disease Control and Prevention, since 1999, it is usually first found in sick poultry and then diagnosed in livestock and humans.

In addition, Kelly said, “We can detect the first invasive species animal that appeared in the center of California.” For example, if the number of mourning pigeons entering the wildlife center suddenly changes, the system can issue an alert and issue an Eurasian collared pigeon to the veterinarian. Signals that have arrived; they are an invasive species that compete for food and spread parasites to native pigeons.


Source link