The research explores the application of AI-driven sequence data analysis in two healthcare areas: blood group antigen characterization and enhancing intelligent upper limb prosthetics. The first study explores how AI models can improve our understanding of blood group antigens. The goal is to identify and understand changes in antigenicity caused by mutations to reduce adverse reactions during blood transfusions, thereby improving maternal and foetal health outcomes and enhancing healthcare efficiency.
The second objective is centred on improving the reliability and functionality of upper limb prosthetics. This shall be achieved by integrating multimodal data, including electromyographic signals from residual limbs and speech commands. The research aims to create a more intuitive and effective control mechanism for prosthetic devices, ultimately enhancing the user experience and promoting social and workforce inclusion for individuals using prosthetics.
This research seeks to identify cross-domain methodologies and algorithmic synergies to drive innovation in healthcare through generative AI and advanced machine learning models. Combining insights from protein structure prediction and prosthetic control, the research seeks to contribute to personalised medicine and intelligent assistive technologies, pushing forward innovation in healthcare applications.

This dissertation project explores the complex field of sustainable finance, focusing on the impact of EU regulations on the financial sector. It consists of three papers. The first paper employs bibliometric methods to analyze how academic research addresses EU regulations on sustainable finance. It identifies key research clusters and highlights how EU policies shape research agendas. The second paper examines the impact of the Corporate Sustainability Due Diligence Directive (CSDDD) announcement on the stock returns of European companies, analyzing the relationship between ESG practices and stock price resilience. The third paper compares the financial performance and sustainability of investment funds according to Articles 8 and 9 of the SFDR regulation. It investigates whether "dark green" funds (Article 9) exhibit higher financial and sustainability performance than "light green" funds (Article 8).

At the time the Persons and Companies Act (PGR) was created, Liechtenstein was experiencing a challenging economic situation due to the events of the First World War and was in urgent need of capital. Consequently, the country sought to strengthen its economic and legal ties with Switzerland, with a view to establishing a new, innovative company law. In conjunction with a liberal tax law, Liechtenstein's legislators wanted to make the Principality more attractive to foreign investors. The PGR, which came into force in 1926, therefore contains a diverse range of legal forms that are intended to provide every interested investor with a suitable company form or asset structure that is specifically tailored to their needs. With regard to Liechtenstein legal forms for asset structuring, the foundation and the trust should be mentioned in particular. The enduring popularity of these structures is underscored by the latest figures, which show that approximately 9,500 foundations and around 1,600 trusts are currently registered in Liechtenstein.
Following a period of stability, there are now extensive national reform efforts underway to update the rules governing foundations and trusts. These primarily concern control and supervision, as well as the associated intervention in a closed, self-regulating system. On the one hand, the government of the Principality of Liechtenstein adopted a consultation report in November 2023 on the optimisation of trust law, in which far-reaching legal amendments are proposed. The planned changes relate in particular to trust governance. The proposal's key points include the introduction of a mandatory information officer (enforcer), amendments to the provisions on judicial supervision, the catalogue of supervisory measures, the right to file an application and party status in supervisory proceedings, and the subordination of charitable trusteeships to the supervision of the foundation supervisory authority. On the other hand, following the total revision of foundation law that took place in 2008, a targeted improvement of the existing standards is also being considered. Adjustments are to be expected in particular with regard to the beneficiaries' rights of inspection and information, supervision and the prevention of legal disputes between the parties involved in the foundation.
From a comparative legal perspective, the two legal forms of foundation and trust are also subject to regulatory influences at the European level. In this context, the possibilities for the Liechtenstein financial centre to position itself and stand out in international competition must be examined. The objective of the research project is to subject the planned national legislative changes to a scientific analysis, to place them in an international context and, finally, to analyse them in a comparative legal context. In addition, the
European influences on other legal forms of asset structuring, such as corporations as family pools or holding companies, will also be assessed.

As part of a joint research project between the Chair of Sustainable Finance and Investments at the University of Liechtenstein and the prosperity company in Liechtenstein, the potential for integrating impact investing components into equity-based pension provision products was investigated. The project focused on the conceptual framework, regulatory context, and an empirical analysis of sustainable funds (SFDR Articles 8 and 9).
Building on the outcomes of this project, the next phase of research aims to develop a scientific publication that deepens and expands on the key questions identified.

The planned focus areas of this follow-up research include:

• A nuanced analysis of the mechanisms through which sustainable funds generate impact, particularly by comparing ESG metrics and EU taxonomy indicators;
• The development and empirical validation of an evaluation model for identifying high-impact funds, taking into account risk, return, and sustainability dimensions;
• An in-depth examination of regulatory uncertainties and their implications for the classification, credibility, and marketing of impact investing products;
• The simulation and assessment of investor portfolios with varying sustainability prefer-ences under different market scenarios.

This extended research aims to contribute to the academic foundation of impact investing in the context of private pension provision, while also offering practical insights for financial product development and regulatory frameworks.

tbd

Sebastian Petric's doctoral thesis focuses on using machine learning methods to understand, identify, and predict financial crises and their impacts on markets and investment strategies. His research aims to develop systems that not only detect potential financial turbulence but also guide investment strategies. By leveraging machine learning techniques such as predictive modeling, and unsupervised learning techniques, his work seeks to create data-driven approaches that identify and forecast crises and inform more resilient and adaptive investment decisions. The research underscores the critical need to understand the complex dynamics of financial crises to enhance risk management and optimize investment strategies, providing valuable insights for financial experts, policymakers, and the broader public.

Sebastian Petric's doctoral thesis focuses on using machine learning methods to understand, identify, and predict financial crises and their impacts on markets and investment strategies. His research aims to develop systems that not only detect potential financial turbulence but also guide investment strategies. By leveraging machine learning techniques such as predictive modeling, and unsupervised learning techniques, his work seeks to create data-driven approaches that identify and forecast crises and inform more resilient and adaptive investment decisions. The research underscores the critical need to understand the complex dynamics of financial crises to enhance risk management and optimize investment strategies, providing valuable insights for financial experts, policymakers, and the broader public.

The decline in church attendance in Switzerland mirrors broader secularization trends in Europe, leading to an increasing number of underutilized ecclesiastical buildings, many of which belong to the so-called years of the postwar period. Despite decreasing attendance, a certain degree of uncertainty in church engagement persists as events like Christmas and crises temporarily boost participation. As most ecclesiastical buildings were not designed in a flexible way, and since preserving them is more sustainable than demolition, repurposing them in a flexible adaptive way offers a more viable solution, optimizing costs and aligning with circular economy principles. On top, most of these churches are in priviledged locations, well served by public transport and often in the centers of cities, where their presence is both historically, socially and architecturally significant, making their future use a matter of urban interest. This creates a crucial sustainability challenge, as these structures still require substantial resources despite being only partially occupied. Conventional repurposing strategies often involve costly renovations with limited flexibility, failing to align with principles of flexible adaptive repurposing. Furthermore, the carbon footprint of demolition and new construction makes it imperative to find innovative ways to repurpose religious spaces minimizing the environmental impact.
Despite recent progress, significant gaps remain in research on sustainable adaptive repurposing and long-term resilience strategies. These in particular, include the lack of a theoretical foundation for integrating cutting-edge computational and economic modeling techniques, as well as the absence of real-world operationalization of these tools to optimize flexibility strategies. Generative Artificial Intelligence (GenAI) and Foundation Models (FMs) — which enable automated scenario generation and predictive design processes — have not been systematically applied to architectural repurposing of the ecclesiastical cultural heritage. At the same time, Real Options Theory (ROT), a financial decision-making framework that quantifies investment flexibility under uncertainty, remains largely unexplored in the context of adaptive ecclesiastical repurposing. Current methodologies do not fully account for future uncertainties in demand for religious spaces or provide a scalable framework balancing economic feasibility, social utility, and environmental sustainability.
This project, Machine-learning for Adaptive Resilient Intelligent Architecture (MARIA), proposes a novel AI-driven framework that integrates GenAI, FMs, and ROT to enable the flexible, sustainable repurposing of underutilized church buildings in Switzerland. The methodology is structured into three phases: MARIA applies ROT to AI-driven decision-making, first validating the framework through a single case study — the church of St. Felix and Regula in Zurich — before extending it to a broader dataset of ecclesiastical buildings from postwar modernism of the 1950s and 1960s which are often affected by the issue of repurposing and flexible use. Finally, it develops a decision-support tool for adaptive repurposing applications beyond religious structures. By incorporating GenAI-driven simulations, FMs, ROT-based flexibility quantification, and augmented data from the Schweizer Kirchenbautag database, MARIA provides an innovative approach to preserving historical ecclesiastical value through flexible sustainable adaptive strategies. The MARIA project pioneers the integration of GenAI, FMs, and ROT into sustainable adaptive repurposing by evaluating the impact of flexible designs for postwar modernist churches while coping with the uncertainty in future demand.

We are investigating how artificial intelligence algorithms can support musicians in making claims for music royalties that have been lost in the confusing process of digital processing. Another project covers Generative AI and dance, where we want to use modern deep learning architectures to develop tools for dancers that enable them to investigate the creative potential of artificial intelligence and make it fruitful for their creative process.

Introduction and Background
The Federal Reserve (Fed) significantly influences U.S. economic policy. Fed speeches, pivotal for market participants, provide insights into monetary policy, economic forecasts, and regulations. Over time, the Fed has shifted towards transparent communication to enhance policy effectiveness. Analyzing these speeches with machine learning (ML) and natural language processing (NLP) offers valuable insights into their impact on financial stability.
Research Objectives
This research aims to automate Fed speech classification and measure their effects on financial markets. Key questions include:
1. How can ML classify Fed speeches by relevance to financial stability?
2. What are the short-term and long-term market impacts of these speeches?
3. How can we distinguish between anticipated and actual speech effects?
Methodology Data Collection and Preparation
A comprehensive database of Fed speeches will be created, preprocessed, and enriched with metadata like date, speaker, and context.
Text Classification Using Machine Learning
Various ML and NLP techniques, including Bag-of-Words, TF-IDF, Word Embeddings, and Transformer Models (e.g., BERT, GPT-3), will classify the speeches.
Model Training and Evaluation
Using annotated data, models such as SVMs, random forests, and neural networks will be trained and evaluated with accuracy, precision, recall, and F1-score metrics.
Analysis of Impact on Financial Markets
The classified speeches' impact on stock prices, bond yields, and exchange rates will be analyzed using statistical and econometric methods, including event studies and time-series analysis. The goal is to separate anticipated market effects from actual impacts post-speech.
Relevant Literature and Theoretical Foundations
This research builds on the work of Jurafsky and Martin (2020) on NLP, Goldberg (2017) on neural networks, Blinder et al. (2008) on central bank communication, and Hansen and McMahon (2016) on transparency.
Expected Outcomes and Implications
The classification method is expected to differentiate speeches by their relevance to financial stability reliably. Analyzing speech impacts will provide insights into monetary policy communication. This research could enhance Fed communication strategies and contribute to financial stability. It will offer a real-time tool for market participants and policymakers, aiding in risk management and decision-making.
Conclusion
This interdisciplinary approach combines ML with economic analysis to study Fed speeches' impact on financial markets. The anticipated outcomes will advance research and practical applications in financial economics, enhancing central bank communication strategies.