AI at CORBA

Artificial intelligence (AI) plays a crucial role in the CORBA effort to improve decision-making. The core AI concept of Machine Learning (ML) enables us to enhance decision-making by transforming hidden patterns and dependencies into valuable information.

We leverage ever-growing data across various application domains, such as transportation, finance, and manufacturing. We then create predictive ML systems that support decision-makers or deploy automated decision systems. Particularly when traditional approaches are not feasible or efficient, we use cutting-edge Deep Learning methods for predicting or developing autonomous agents through deep reinforcement learning techniques.

In addition to enhancing the quality of decision-making, we also prioritize gaining acceptance of these innovative systems by stakeholder who either use these systems or are affected by them. We address concerns such as explainability and fairness through behavioral experimental studies. This allows us to explore ways to alleviate issues like algorithm aversion or loss of control during the design phase of AI-based systems.

AI-based Decision Support

We develop AI-based systems that transform complex, multimodal data into valuable insights - supporting decision-makers with predictive and prescriptive models.

AI-driven Decision Making

We combine AI and mathematical optimization to support real-time operational decisions, from pricing and demand management to fleet allocation, inventory, and routing.

Human-AI Interaction

We design AI systems that are not only technically effective but also trusted, fair, and understandable, ensuring they support users and positively impact work and society.

Prediction Models

We develop AI-based prediction tools for urban delivery and mobility, helping companies optimize operations and providing accurate, reliable information for customers.

AI-based Valuation and Exercise of American Options

We apply AI and machine learning to price and exercise American-style options, using Reinforcement Learning and Random Forests to tackle complex, high-dimensional financial challenges.