JADBio, an AutoML platform for life-scientists, announced the addition of image analysis capabilities in its platform API interface. The addition of feature construction optimisation using pre-trained Deep Neural Networks for image analysis will allow researchers, clinicians, bioinformaticians, biologists, and biomedical staff to use their medical image datasets to train and deploy machine learning models, without writing any code.
Ioannis Tsamardinos, CEO at JADBio, said, ‘Doing modelling with images used to demand Deep Learning experts, thousands of images, and sizable computational resources; it is error-prone and easy to overfit. In JADBio, we tackled this problem, did our research and fully automated the procedure. No Deep Learning or Machine Learning expertise is required, statistical correctness is guaranteed, and we have got amazing results even with very few samples, less than 30 images in some cases. Analysing images fits seamlessly with the rest of the JADBio pipeline so one could analyse X-ray images in combination with millions of genetic markers and hundreds of clinical quantities.’
Also Read: Top 10 AI Research Labs Worldwide
JADBio’s AutoML platform enables bioinformaticians, oncologists, virologists, data analysts, and other life-scientists, to build machine learning models with no coding needed on their part. Its platform is intuitive and easy to use, permitting anyone to upload their data securely, train, and develop a model in a matter of minutes, using highly sophisticated algorithms. JADBio can analyse small sample sizes or very large feature sets, focusing on feature selection and interpretation of the predictive model. Its survival analysis functionality helps clinicians predict life expectancy, cancer metastasis and solve any time-to-event prediction problem. The platform comes with an array of reports and visual graphs to help interpret the model.
