Case Studies on AI System Accuracy

In this study, AI was more accurate than two thirds of radiologists, yet when radiologists had AI help their diagnoses did not improve. Why? Humans ignored the AI’s advice when it conflicted with their views. A big barrier to future human-AI collaboration. In fact, the paper concludes: “Our results demonstrate that, unless the documented mistakes can be corrected, the optimal solution involves assigning cases either to humans or to AI, but rarely to a human assisted by AI.” Paper: https://lnkd.in/eKGBtEAk

Fantastic case study from Cincinnati Children's Hospital on Identifying #mentalhealth concerns, subtypes, temporal patterns, and differential risks among children with #CerebralPalsy using #NLP on #EHR data. Adding NLP features improved accuracy by 15% compared to only using structured data. In this use case, accuracy can imply earlier diagnosis, which has material clinical outcome benefits to children. https://lnkd.in/gY9UUDbG #ai #generativeai #healthcareai #healthai #nlproc #textmining #medicaldata John Snow Labs

A great example of using Artificial intelligence (AI) and Precision Medicine to revolutionize Parkinson's diagnosis. The complexities of Parkinson's disease puzzle for ages, AI strides in, illuminating the way. They reported that machine learning models achieved an accuracy of 95% in predicting #Parkinson's disease presence and accurately classified the subtype of Parkinson’s disease in their paper in Nature Machine Intelligence https://rdcu.be/djgwp. Using patient-derived induced pluripotent stem cells (#iPSC) and chemically creating four different four subtypes of #Parkinsondisease, they created a ‘human model of brain diseases in a dish’; two pathways leading to a toxic build-up of α-synuclein (α-Syn ) protein and two leading to faulty mitochondria. The Author, James Evans, a Ph.D. student The Francis Crick Institute and UCL “An unmet challenge is to make an early and accurate molecular-level diagnosis of the condition, as this would enable the field to consider targeted interventions appropriate to an individual’s condition, and offer an opportunity to do this at the earliest possible time.” @James Fleming, PhD chief information officer at the Crick who worked with Faculty AI on the project, said, “AI is a fascinating and powerful technology, but one which is often rendered impenetrable by hype and jargon. ………..see if a group of complete AI beginners could learn and apply best practices directly to their science in a very compressed time frame. The success of this project not only proved that they could, unlocking new insights in the process but has also helped drive investment into the rapid expansion of our own AI and software engineering team, which has over 25 projects ‘in-flight’ with different labs across the Crick, with new projects kicking off every month.” Variability in Parkinson's disease and its symptoms is due to heterogenicity and various underlying mechanisms. The study found that #protein misfolding, dysfunctional #mitochondria, and #lysosomes were the most essential features in predicting the correct subtype. The study's success could pave the way for #PersonalizedMedicine strategies and #targetedtherapy in Parkinson's disease treatment.  In the grand tapestry of scientific progress, AI's synergy with healthcare is stitching a masterpiece. #AIandHealthcare #ParkinsonsPrecision #ScientificInnovation #MedicalProgress #medicalinnovation