AI models often find it challenging to interpret the captions of certain medical images
Getty Images
Although young children may quickly grasp the concept of “no,” numerous artificial intelligence models struggle with this. They demonstrate a significant failure rate when interpreting commands that include negative words such as “no” and “not.”
This could imply that we may not differentiate effectively between x-ray images marked as “indications of pneumonia” and those marked “no indications of pneumonia.”
Source: www.newscientist.com
Discover more from Mondo News
Subscribe to get the latest posts sent to your email.