Impara (Intelligent Machines for Pattern Recognition and Adaptive modeling), is a spin-off of the National Research Council of Italy founded in 2007. Its primary mission concerns the design, the development, the implementation and the provision of statistical and machine learning methods. Impara develops the core algorithms of Rulex, an integrated suite for extracting knowledge from data through classical and advanced techniques. Besides standard algorithms, such as decision trees, neural networks, logistic, and k-nearest-neighbor, Rulex offers the possibility of applying an original proprietary approach, named Logic Learning Machine (LLM), that allows to produce sets of intelligible rules capable of achieving an accuracy comparable or superior to that of best machine learning methods. Currently, Impara employs Rulex in different R&D projects to analyze data coming from a variety of areas, including health, communications, marketing, energy.
In the communication field, Impara developed for national projects several solutions involving: QoS control, inference from sensors and location. In the last years, Impara has been developing a new kind of adaptive control, based on LLM, which can be employed even in case of highly complex systems, where an explicit or approximate model is not available.
- Andrea Ridi
- Enrico Ferrari
- Massimiliano Costacurta