(1) RegCheck: Program analysis for regulatory compliance assessment of FinTech Software [March 2023 – February 2025] RegCheck aims to derive privacy rules from GDPR and verify whether software artifacts (e.g., source codes) of FinTech applications (e.g., online banking) properly implement such rules to be compliant with GDPR. Our work within ARTAGO project is relevant and complementary to RegCheck. Drawing on that, we contribute to the elicitation and representation of privacy rules from GDPR.
(2) AFRICA: Automated Financial Regulations change Impact Analysis [May 2023 – June 2025] AFRICA aims to (semi-)automatically detect the change in the landscape of financial regulations and analyze the impact of this change on the compliance process, notably in various software artifacts (e.g., NL requirements) of financial applications. The project will further provide recommendations to address the detected changes and remain compliant. We contribute to the development of the automated solutions in AFRICA.
(3) RUMOFA: Runtime monitoring of fund activities [May 2023 – June 2025] RUMOFA aims to automatically monitor and verify the financial transaction records against the financial requirements. The project will rely on NLP technologies to process and extract meaningful information from the financial regulations and fund documents to enable verifying fund activities in real-time. We investigate innovative methods to transform the extracted information into some intermediate format that can be used in the runtime monitoring.
(4) ICCOFIDO: Incremental compliance checking of financial documents activities [May 2023 – June 2025] ICCOFIDO aims to detect the temporal changes in financial documents (prospectuses) which are likely to impact the compliance process. ICCOFIDO will develop automated means for checking compliance incrementally, i.e., accounting for the changes across multiple revisions of the same document and further assessing these changes against the compliance rules derived from the financial regulations. We contribute to developing automated solutions for detecting temporal changes in prospectuses and analyzing the detected change on their compliance.
PLAITO: Automated Completeness Enhancement of Requirements towards Improved Trustworthiness [September 2024 - August 2027] Artificial intelligence (AI) is transforming business and industry, including software development practices. To foster the economic and social benefits of AI, regulations are being enforced to ensure that the development, deployment, and use of AI is trustworthy, i.e., ethical, lawful, and robust. Trustworthiness in modern software systems (e.g., AI-enabled software) should be addressed in the upstream stages of software development, mainly requirements engineering (RE). RE is concerned with introducing quality assurance measures to ensure that the software qualifies as trustworthy. With modern software becoming more complex, traditional techniques for quality assurance in RE should be revisited and revamped. PLAITO provides a new perspective on enhancing the completeness of natural-language (NL) requirements, with an emphasis on diversity and non-discrimination ethical concerns. To detect incompleteness violations, PLAITO will put forward a technical approach that employs recent artificial intelligence technologies, including natural language processing (NLP) and machine learning (ML). More concretely, PLAITO investigates the powerful capability of large-scale language models (LLMs) in language understanding and generation. Building on the extensive knowledge of LLMs, PLAITO will develop automated solutions that can expand the current view of engineers and pinpoint the potentially missing content in the existing requirements. PLAITO will investigate the possibility of utilizing LLMs for automatically generating user stories that describe diverse users and their needs. The automatically generated user stories will be used to detect incomplete and/or missing requirements which should capture diverse users’ needs. PLAITO will further complement the research agenda of responsible NLP by empirically evaluating both the effectiveness of several LLMs in generating in-domain textual requirements, and the sustainability of LLMs in terms of their energy consumption. A successful completion of PLAITO will present an important academic advance in the field of requirements engineering and at the same time provide several avenues for industrial exploitation.