Scientists from Drexel University's School of Biomedical Engineering, Science and Health Systems
Traditionally for the diagnosis of this diseaseanalysis of the patient's medical history and a series of physical and neurological examinations and tests are used. Researchers have focused on programs that can pick up subtle cues, such as hesitation, grammatical and spelling errors, and forgetting the meaning of words, as a quick test that can indicate whether a patient should undergo a full examination.
To diagnose Alzheimer's disease scientistsretrained the GPT-3 AI based on a subset of the patient voice recording dataset. A computer algorithm extracted meaningful word usage, sentence structure, and semantic characteristics from text to create what scientists call "embedding" -- a distinctive speech profile of Alzheimer's disease.
To test, the researchers usedan additional test set of data collected from a control group of patients. The scientists asked the program to evaluate dozens of transcripts and determine whether each of them was written by a person suffering from Alzheimer's disease or not.
The results show that the embedding scoreThe text obtained by GPT-3 can be reliably used not only to identify people with Alzheimer's disease among healthy people, but also to assess the patient's cognitive test, the scientists said. The researchers plan to create a web application that can be used at home or in the doctor's office as a pre-screening tool.
Third version of General Pretrained Transformer(GPT) uses deep learning to create human-like text. Using the source material as a clue, he creates text that continues it, paying special attention to how words are used and how language is built.
The tomb of the “midwife of Jesus” was unearthed: scientists told what they found there
Einstein is wrong again and his main theory was rewritten: how it changes the world
Published test video of the world's first propeller with 11 blades