The study, published in the journal Modern Pathology, is the first to examine the impact of tissue contamination on machine learning models.
The researchers trained three AI models to scan microscopic slides of placental tissue to detect blood vessel damage, estimate gestational age, and classify macroscopic damage. They also trained a fourth AI model to detect prostate cancer in tissue collected via needle biopsy.
It was found that when AI models were exposed to small amounts of contaminated tissue, each model paid too much attention to the contamination, leading to diagnostic errors. According to the researchers, this is because AI models are often trained in pure, simulated environments and are not used to dealing with real-life distractions.
The study’s findings have important implications for the use of artificial intelligence in pathology. Pathologists rely on AI to make diagnoses, but research suggests that AI models should be used with caution, especially when working with slides that may be contaminated.
Researchers say one way to solve this problem is to train AI models on a broader range of data, including data containing tissue contaminants. They also suggest that AI models can be designed to be more resilient to distractions.
Source: Ferra
I am a professional journalist and content creator with extensive experience writing for news websites. I currently work as an author at Gadget Onus, where I specialize in covering hot news topics. My written pieces have been published on some of the biggest media outlets around the world, including The Guardian and BBC News.
