Context and General Objectives 

The digital age demands tools for intelligent data analysis. Artelnics developed Neural Designer, a deep learning solution for predictive analytics, to help organizations make better decisions from their vast data. It’s successfully applied in engineering, marketing, and health. 

The project’s main goal was to integrate an innovative model selection framework into Neural Designer to handle complex datasets, requiring support for big data and super-computing clusters. Work Performed and Key Results During Phase 1 of the SME Instrument, Artelnics made substantial progress across the technical, commercial, legal, and financial dimensions of Neural Designer. 

On the technical front, the development team has expanded, and a new, more optimized version of Neural Designer, packed with innovations, has been released. Crucially, various model selection algorithms have been developed to automatically identify the most relevant variables within a dataset. 

Advances Beyond State-of-the-Art and Expected Impact 

Many applications of predictive analytics, ranging from customer segmentation to medical diagnosis, stem from complex interactions among diverse variables. A software tool that can efficiently untangle these factors represents a significant innovation with the potential to disrupt the existing predictive analytics market. Neural Designer is poised to overcome these challenges, guaranteeing optimal results by offering innovative model selection techniques coupled with high computational performance. The immediate benefit for customers is a strengthening of their businesses, enhancing their competitiveness. 

The potential societal impact of applying predictive analytics is vast; in engineering, it can lead to reduced gas emissions or minimized equipment faults; in business, it aligns products with customers and improves the buying experience; and in health, data mining can contribute to better diagnoses and more effective treatments. 

Neural Designer achieves this by seamlessly integrating the best techniques from artificial intelligence and computer engineering to construct the most powerful predictive models from large-scale datasets, thereby enabling the development of truly impactful predictive analytics applications.