Machine Learning
Machine Learning is the analytical engine behind CAOS. We develop algorithms that enable computers to identify hidden patterns and learn from experience without being explicitly programmed. Specializing in intelligent data analysis, we create highly robust predictive models capable of operating in industrial environments with high levels of uncertainty. Examples of this include Multiple Classification Systems (Ensembles) and Time Series Analysis and Forecasting.
PredicciĂłn de Series Temporales
Using Neural Networks to Predict Financial Market Behavior and Energy Demand.
Reconocimiento de Patrones
Classification algorithms for identifying human behavior and complex operational flows.
Conjuntos de Clasificadores
Combining multiple learning models to maximize the accuracy of data diagnosis.
Laboratory Milestones
Multiple Classification Systems (Ensembles): Much of their success in computer vision and prediction is based on not using a single algorithm, but rather on training multiple algorithmic “experts” that vote to reach a final decision (such as Random Forests or Boosting). This drastically reduces false positives.
Time Series Analysis and Forecasting: Neural networks are used to predict future behavior based on historical data. This technology has been applied in businesses to forecast product demand, anticipate industrial machinery failures (predictive maintenance), or detect financial anomalies.
Intelligent Decision-Support Systems: Technological solutions based on autonomous artificial intelligence agents that analyze vehicle, environmental, and physiological data—among others—in real time. These systems assist human drivers and autonomous vehicles with navigation, hazard avoidance, and route calculation, which ultimately optimizes traffic flow and improves safety.
Strategic Consortia: Our group has been a key player in programs such as RedAF (Spanish Thematic Network on Agent Technology and Multi-Agent Systems), an open network of researchers that serves as a forum for communication on research, innovation, and technology transfer in the field of physical agents, encompassing areas such as embedded systems, artificial intelligence, mobile robotics, home automation, and ubiquitous systems.
«At times, the only way to be right is to be wrong.»
– JosĂ© BergamĂn –