With visualisations of feature importance, ICE-plots, and analysis results the user is able to gain insights and knowledge about the models and their data.
AutoML has been used for all sorts of ML problems ranging from estimating biological and disease monitoring age to autism screening and suicide risk assesment. Still, the platform is tailor made for bioscience featuring bio-specific functionalities such as Survival analysis and bio-specific data preprocessing and is able to handle complex problems simultaniously handling different data types such as multi-omics, genetic data, images, and medical signals.
The platform can be accessed through the REST API, allowing power users to leverage more advanced functionality and incorporate it into company software and pipelines.
The REST API is accessible through a simple Python client, making it easy to use AutoML with and alongside the cornucopia of life science modules provided for Python.
Raven biosciences are currently building a cheminformatics platform meant to add strong cheminformatics functionality to JADBio’s AutoML. It is developed along the same philosophy as JADBio – making machine learning easy and accessible whilst still providing power users deeper control over the underlying processes to accomplish specific tasks.
The software will offer the user a suite of options for data wrangling, featurization, analysis, deployment of model, and knowledge extraction all useful for focusing drug discovery efforts.
The software will offer its users state-of-the-art ML modelling with added, chemistry specific functionalities for knowledge extraction and visualisation.