2 min

AI company Aible wants to support automated machine learning (autoML) for data scientists and developers. The Aible Advanced solution should provide automated generation of predictive models for companies.

The solution must include all the steps of autoML. Aible states that not all steps of automated machine learning are automated by legacy autoML solutions. According to Aible, an autoML process starts with conceptualizing objectives and then viewing blueprints of sample models. Subsequently, ‘data receptacles’ must be created for the collection of relevant data, after which data cleansing will take place. Model customization is the next step, to allow a model to deliver maximum performance for a specific business objective. Then comes hyperparameter tuning, training multiple models to choose the best from the group. The selection of a model follows, after which the model is rolled out in the business environment. The last two steps are prediction writeback, which involves communicating forecasts to business applications, and monitoring the model to evaluate how well it is performing.

Save time and effort

The solution must include all these steps. According to Arijit Sengupta, founder and CEO of Aible, the solution will enable business users and data scientists to work better together: “Business users have the knowledge of the business domain; data scientists have the modelling expertise. Aible Advanced, in collaboration with Aible, allows each type of user to seamlessly contribute his or her unique skills and knowledge to generate the best prediction model for his or her unique business reality.”

The solution should save developers a lot of time by automating repetitive tasks. Aible also takes on a kind of challenge with potential users. With data from a company, Aible Advanced then develops a model in 30 minutes. If the model does not do what a company asks, the company receives the model in question for free. If it does, the company will pay for the model.

This news article was automatically translated from Dutch to give Techzine.eu a head start. All news articles after September 1, 2019 are written in native English and NOT translated. All our background stories are written in native English as well. For more information read our launch article.