Brainome, the company behind the measure-before-build tool for machine learning, has launched Daimensions for the enterprise market.
By introducing measurement as a structured discipline to the field, Brainome claims to change the way organizations approach machine learning: faster, more effective data preparation, rapid training and execution of models, compact model size, quick identification of most predictive data attributes, and explainability of results.
Brainome ultimately delivers what it claims to be a new and complete structured method for analyzing data and creating models.
Brainome’s systematic measurement-based approach is aimed at steering teams towards better outcomes, making it possible to predict project speed and costs.
By measuring first, customers can know whether there’s enough data to learn rules and avoid overfitting, and find and design the data features that matter.
They can also iterate through data prep and model design options without training, and train and execute rapidly using compact, efficient models.
Brainome was co-founded by former business partners Gerald Friedland and Bertrand Irissou in Berkeley, California.