ºÚÁϸ£ÀûÍø

Keynote Francisco Camara Pereira

Climate change is expected to increase the frequency and intensity of extreme rainfall, leading to more frequent urban floods that damage infrastructure and disrupt mobility. Cities urgently need strategies to adapt to these escalating risks. In this seminar, I will present how reinforcement learning (RL) can be used to uncover effective adaptation strategies—identifying where and when measures should be deployed. The framework combines projections of future rainfall and flooding with a simplified model of Copenhagen transport, allowing us to capture both direct damage to infrastructure and indirect impacts on mobility. I will share preliminary results showing how RL can help prioritize interventions in vulnerable areas and determine the optimal timing for their implementation, ultimately supporting more resilient and adaptive urban transport systems.

Karolin Winter

In this talk, I will argue that we need new ways of thinking about errors and analysing them, focused on understanding how users perceive them, and how factors such as context or emotions affect lived experiences with errors. I will present a vocabulary we created as a first step towards designing human-data interaction experiences that can mitigate tensions emerging from encountering data and AI errors.

Irina Tentina

Process mining enables organizations to gain insights into their business processes using event logs from IT systems and applications. Some organizations manage to achieve valuable results from process mining, while others still struggle with its adoption or even end their process mining initiatives. Many factors influencing the adoption of process mining have been studied recently, such as lack of management support or poor data quality. However, little attention has been paid to business users’ trust in process mining analytics.

In this study, we explored how business users assess their overall trust and confidence in process mining implemented within their organization. We report on the results obtained from interviews with process mining users from various industries. I will present common trust-building approaches followed by business users and challenges they face when validating the quality of process mining output. Then will share practical implications for practitioners that can help build trust into process mining.

Alessandro Corbetta

Accurately predicting how crowds move through large public infrastructures is an outstanding physics challenge, essential for ensuring both safety and comfort — from the daily flows of commuters in busy train stations to maintaining security at major events.

In this talk, I will present our latest data‑driven models, developed from large‑scale real‑world pedestrian tracking. These models not only capture movement patterns within complex public facilities and provide actionable insights for design, and crowd management, but also uncover the underlying structure of pedestrian-pedestrian interactions.

Raquel Salcedo Gil

This Ph.D project investigates how employees adapt to working with robots in industrial environments, with a focus on the psychological, behavioral, and work design implications of Human-Robot Collaboration. It examines how the introduction of robots transforms job design—creating new knowledge and skill demands—and explores how organizations can effectively support employees in navigating this transition.

Shuxia Tao

At the core of my research is a simple but ambitious question: How do electrons, ions, and atoms interact to control energy conversion, storage and transport, and how can we predict and design these processes using AI? In my group at ºÚÁϸ£ÀûÍø, we combine deep physical understanding with machine learning to build predictive models for complex materials behavior. This takes us from simply observing, to truly designing new materials with purpose. AI doesn’t replace physics, it unlocks its full predictive power. By bridging quantum mechanics to real-world devices, we aim to accelerate the discovery of next-generation energy materials.