AI, NLP & Financial AI Research Projects
My research projects span Financial AI, multilingual NLP, trustworthy AI, low-resource language technologies, document intelligence, generative AI, and large language models. These projects involve collaborations with universities, governments, regulators, and industry partners across the UK, Europe, the Middle East, and Southeast Asia. Many of the projects focus on building scalable AI infrastructure, multilingual datasets, explainable AI systems, and domain-specific language technologies for high-stakes environments including finance, policy, healthcare, and sustainability. :contentReference[oaicite:0]{index=0}
Major & Recent Projects
PolicyVerse
A multilingual multi-agent LLM framework for policy reasoning and collaborative multilingual reasoning (CMR). The project focuses on detecting policy overlaps, inconsistencies, and dependencies across multilingual policy documents using advanced AI reasoning systems.
Dynamic Models for Real-Time Autonomous Decision Making
Research on autonomous decision-making systems under uncertainty, including stochastic environments, legacy teleoperated systems, and continuous action spaces for real-world AI deployment.
Detecting Greenwashing and Greenhushing Risks
An NLP-driven framework for identifying greenwashing and greenhushing risks in corporate sustainability disclosures to support responsible finance and ESG transparency.
FreeTxt-Vi
A bilingual Vietnamese-English free-text survey analysis platform for automated processing, summarisation, sentiment analysis, and large-scale analysis of open-ended responses.
ViSAS
An open-source Vietnamese semantic tagging system inspired by the USAS and AraSAS semantic analysis frameworks for multilingual semantic understanding.
Monitoring Water Pipe Bursts with NLP
Using natural language reports and operational text data to predict and monitor water infrastructure failures using applied NLP and machine learning techniques.
Welsh Language Model Pilot
A pilot transformer-based language model for Welsh, focusing on sentiment analysis and low-resource language AI development.
DigiGrid
Infrastructure for collecting, managing, and curating Welsh language resources to support NLP and AI research for low-resource languages.
Catalyst for Celtic Languages NLP Research
Supporting the development of datasets, NLP tools, and AI resources for under-resourced Celtic languages.
FreeTxt (UK)
A framework for processing bilingual survey and questionnaire free-text data using NLP, corpus linguistics, and explainable AI techniques.
Earlier Projects
CARE – Canadian Annual Reports Extractor
Development of datasets and pipelines for analysing Canadian financial disclosures using NLP and document intelligence techniques.
Beyond the Paycheck
Research collaboration focusing on workplace wellbeing, organisational communication, and large-scale textual analysis.
Thesaurus of Contemporary Welsh
Development of a modern Welsh thesaurus using embedding-based semantic modelling and NLP approaches for low-resource language technology.
Welsh Summary Creator
Automatic summarisation system for Welsh language texts using transformer-based NLP and multilingual summarisation approaches.
CLARA-Fin
Research on readability metrics and textual complexity in English and Spanish financial disclosures using NLP and corpus-based methods.
Corporate Disclosure Assessment (AS15)
Evaluation and analysis of corporate narrative disclosures and IFRS-related reporting using NLP-driven methodologies.
AraSAS
An Arabic semantic tagging framework aligned with the USAS semantic analysis system to support multilingual semantic understanding and Arabic NLP research.
FinT-esp
Building and analysing Spanish financial corpora for multilingual financial NLP and AI research.
Bio Text Mining (BioTM)
Biomedical text mining research focused on mental health and biomedical literature analysis using AI and NLP methodologies.
Influences on Corporate Financial Information Environment
Research investigating relationships between corporate reporting, media narratives, and investor response using computational linguistic analysis.
Understanding Corporate Communication
Analysis of strategic narrative communication and corporate disclosure practices using NLP and corpus-driven methodologies.
SKOS-HASSET
Development of a SKOS-based thesaurus and semantic infrastructure for social science terminology and intelligent information retrieval.