Rapporteur Review
Independent External Assessment of Progress at 24 Months (14/09/2020 - 14/09/2022)
Rapporteur's Executive Summary
Major Outcomes, Impacts and Successes
"The FinTech and AI in Finance Action has demonstrably been highly successful in nurturing the field, and supporting the development of a strong community of data scientists, economic and social science experts and practitioners, and a wide array of stakeholders, including regulators and policy-makers from 49 countries."
The programme has made impressive progress throughout its first 24 months, generating over 100 peer-reviewed publications, numerous cross-national collaborative initiatives, and supporting a sizeable cohort of early career and other academic researchers.
The Action has supported the development of a vibrant and diverse scientific community through a significant number of initiatives, including academic and practitioner workshops, conferences, meetups, 'datathons' and other activities aiming to engage researchers and practitioners associated with FinTech.
Key Rapporteur Findings
Operating within a highly challenging field, the COST Action FinTech and AI in Finance appears to already play a significant role in promoting the state of the art in financial data science and actively support developments in Explainable and Interpretable AI, a field which is crucially important in financial services, especially with reference to end-user and citizen privacy, personal data custody and associated regulatory policy developments in Europe and beyond.
Critical Significance Highlighted
Regulatory and Policy Impact
Of critical significance is the Action's contribution to involving, informing and educating the wider stakeholder population, especially regulators and policy decision-makers around many nascent but highly impactful technological developments in DeFi, cryptocurrencies, decentralised assets and related innovations.
Long-term Platform Contributions
Of similar importance has been the Action's contribution in establishing frameworks and platforms for hosting and sharing datasets for use-case development and testing, a contribution of very high impact, particularly due to the longevity it introduces.
MoU Objectives Assessment
Objective-by-Objective Confirmation
| MoU Objective | Reported | Rapporteur |
|---|---|---|
| 1. Blended approaches for financial services evaluation (ML methods, fraud detection, early warning) | 76-100% | CONFIRMED |
| 2. Black-box model transparency and interpretable/explainable models | 76-100% | CONFIRMED |
| 3. Regulator and practitioner input on AI transparency | 51-75% | CONFIRMED |
| 4. Performance attribution models - reducing false discovery rate | 26-50% | CONFIRMED |
| 5. Dissemination to public and regulators on investment performance | 76-100% | CONFIRMED |
| 6. European platform for investment product performance comparison | 51-75% | NOT CONFIRMED* |
| 7. Excellent research network with lasting collaboration | 76-100% | CONFIRMED |
| 8. Interdisciplinary research in blockchain and FinTech communities | 76-100% | CONFIRMED |
| 9. Knowledge exchange platform bridging industry, academia, regulators | 76-100% | CONFIRMED |
| 10. Knowledge transfer across disciplines and academia-industry | 76-100% | CONFIRMED |
| 11. Inclusive ML/AI research community, ECI promotion | 76-100% | CONFIRMED |
| 12. Geographical/demographical diversity, ITC focus | 76-100% | CONFIRMED |
| 13. STSM and joint educational programs with industry | 76-100% | CONFIRMED |
| 14. PhD and ECI job opportunities maximization | 76-100% | CONFIRMED |
| 15. Dissemination to scientific community and general public | 76-100% | CONFIRMED |
| 16. Gender equality improvement | 76-100% | CONFIRMED |
*Note on Objective 6
Rapporteur Explanation: "Despite providing a reasoned level of alignment, the team is proposing a new objective in replacement of the original set in the MOU; a decision as to whether the new objective is suitable or not does not fully fall into the reviewer's remit. Relevant funder agreement will need to be sought to ascertain progress. Having said that, the reported progress and documentation does fall in line with the newly proposed objective in both scope and level of achievement (51-75% as reported)."
Action Response: Due to the changing regulatory landscape in Europe, enhanced privacy, data protection and regulatory requirements, the original objective is no longer feasible. A replacement objective with substantially higher impact has been proposed.
Publications Assessment
Rapporteur's Publication Assessment
Quality: "Based on a review of the significance and requirements of publication venues, the topics and methodologies covered and publication citation indices, the vast majority of peer-reviewed outputs provided here are of significant quality and indicate concrete contribution to the associated fields."
Significance: "A review on the topics covered, novelty of approaches outlined and relevant citation metrics indicate high significance of the majority of peer-review publications."
Relevance: "The 100 publications range suitably across the entire thematic spectrum of the Action and address highly relevant topics in the FinTech AI space."
Quantity: "There are 100 peer-reviewed publications as outputs of the Action, which rates exceptionally high. This is also a good reflection of the collaborative nature of the work, as the vast majority is multi-author (3 or more academics)."
Projects and Proposals Assessment
The Action reported 10 projects and 10 proposals resulting from Action networking. Notable projects include:
- Crypto Currencies and Nordic Youth (Nasdaq collaboration)
- Network-based Credit Risk Models on P2P Lending (Czech Science Foundation / SNSF)
- Blockchain and Fraud Detection (SNF)
- XAI in Credit Risk Management
- Narrative Digital Finance: Structural Breaks, Bubbles & Market Narratives
Deliverables Assessment
Selected Deliverable Assessments
Database: Crowdfunding/P2P Platform Features
Due: Month 24 | DELIVEREDKey Software (Codes, Packages) by Working Groups
Due: Month 48 | ExpectedDiscussion Papers on ICO/Crowdfunding Methodology
Due: Month 36 | ExpectedMethodological Discussion Paper on AI/ML Stress Tests
Due: Month 48 | ExpectedEdited Volume of Scientific Achievements
Due: Month 48 | Expected within 2 years post-ActionOther Outputs Assessment
"Of the 17 outputs outlined, the vast majority are deemed of high impact, bringing together diverse international audiences and addressing critical novel FinTech topics."
Key Highlight: "Of particular value have been the establishment of databases and hosting of suitable datasets; such activities provide workable platforms for hosting usable data for future research and analysis, guaranteeing long-term impact for the project."
"Similarly, many dissemination activities and presentations have been instrumental in socialising otherwise previously obscure topics (particularly these associated with DeFi) to audiences that are not necessarily technology-savvy, but which are very close to policy and regulation in these areas."
Output Significance Ratings
Key Output Details
TINFIN Conference - Zagreb, May 5-6, 2022
International Scientific Conference "Technology, Innovation and Stability: New Directions in Finance" organized by Faculty of Economics and Business, University of Zagreb, Croatian Academy of Sciences and COST Action FinAI.
- Keynote: "Manifold destiny: Unsupervised machine learning for social sciences" - Prof. James Ming Chen (Michigan State)
- COST session with Action overview by Action Chair Prof. Jorg Osterrieder
- WG2 research presentations and panel on "FinTech in 2022 and beyond"
- 60 participants from 15+ countries
OTP Bank R&D Partnership
Research and development relationship with OTP Bank (largest Hungarian retail bank) on financial services technologies including:
- Blockchain application in financial business processes
- NFT (non-fungible token) research and development
- Digital money, currencies, and CBDCs
- Participation in Hungarian Blockchain Coalition
Blockchain Research Center (BRC)
Founded by COST Action FinAI members Wolfgang Hardle, Vasile Strat, and Alla Petukhina with international researchers at blockchain-research-center.com. Services include:
- Customized independent blockchain solutions
- Scientific support for research and business operations
- High-level academic lectures
- Globally intertwined blockchain forums
Impact Assessments
EIT Digital Summer Schools
Rapporteur Assessment
"The described contribution (summer school) is considered of high S, E and S/T impact, promoting augmented knowledge and expertise in the intersection between finance and data science. The initiative further supports entrepreneurship in the field and nurtures a new generation of tech-savvy Finance professionals. Clear and significant impact is demonstrated."
Virtual Mobility Grants Programme
Rapporteur Assessment
"The VMG initiative is considered of high impact (S/T, E, S) promoting resource and idea exchange across research communities by funding novel ideas and research mobility. There are clear tangible outputs from the initiative so far, also demonstrating the effectiveness of the initiative to date."
Short Term Scientific Missions (STSMs)
Rapporteur Assessment
"Establishment of the referenced grants as parts of the Action are considered of high impact (S/T, E, S) as they promote resource and idea exchange across research communities by funding novel ideas and research mobility. There are clear tangible documented outputs from the initiative so far."
Machine Learning NeEDS Mathematical Optimization Seminar Series
Rapporteur Assessment
"The seminar series is deemed of significant impact across all impact areas to the research community as it clearly supports and promotes a platform for dissemination and research progress on topics which are critical to FinTech areas. Given their purely online nature the seminars appear to support high accessibility for the materials and insights, further attending to the openness and breadth the Action aims to."
Interpretable Machine Learning Workshop (Lorentz Center)
Rapporteur Assessment
"The initiative carries clear and significant impact in buttressing a community of expert researchers around XAI as a critical field for FinTech especially, and in line with regulatory prerogatives in the EU. The extent and focus of the workshop indicates a highly valuable contribution and is considered of high S/T and E significance, while its importance for societal progress is equally high, given the critical role of interpretability and explainability in ML model use."
Crypto Market Research Network
Rapporteur Assessment
"Establishment of diverse research networks is critical for progress in FinTech, especially given the underlying inherent interdisciplinary nature of associated fields. The Action documentation provides ample evidence of sessions organised within such active networks. Furthermore the thematic focus on cryptocurrency research and its support for gender balance in research participants speaks to a crucial societal role in engaging stakeholders holistically. Overall, this is an outcome of high significance in all accounts (S/T, E, S)."
Dissemination Activities
Key Dissemination Events
BlackSeaChain 2022 Conference
Date: September 1-2, 2022 | Location: Varna, Bulgaria
Topics covered: Decentralized economy and finance, regulation challenges, Web 3.0, crypto wallets, NFTs, smart contracts, blockchain/IoT/AI integrations. 43 presentations from academia, industry, and practitioners.
IJCNN 2023 Special Session
Date: June 18-23, 2023 | Topic: "Deep Learning for Financial Data Analysis"
Special session at International Joint Conference on Neural Networks hosted by International Neural Networks Society, featuring presentations on deep learning techniques applied to financial data.
Environmental Finance Workshop
Title: "Environmental Finance for the Common Good"
Co-organized by Money, Macro, and Finance Society (UK) and Birkbeck College, University of London. Featured roundtable on environmental finance challenges, public lectures by Naoyuki Yoshino and Will Goodhart, and awarded best paper by an ECI on fintech/AI topics.
Diversity Workshop - Naples
Title: "Diversity Challenges and Opportunities in FinTech"
Bridging academics and policymakers on FinTech/AI diversity challenges. Featured keynote "AI: Why diversity matters for an inclusive future" by R. Johri (HSBC R&D, Founder of London Women in Machine Learning). Included kickoff for "Woman in FinTech" datathon.
digital-ai-finance.github.io/COST-Fintech-AI-in-Finance/diversity-workshop-naples2022
Lake Como School on Neural Networks
Educational event on neural networks applied to finance providing cutting-edge methodologies. Trained 25 young researchers (PhD students and ECIs). Format included lectures, seminars, hands-on workshops, and networking.
International Fintech Research Conference 2022
Date: October 27-28, 2022 | Host: Politecnico di Milano
35 paper submissions, 24 accepted for presentation. Topics: ML applications, cryptocurrencies, digital currency, cybersecurity, neural networks, smart contracts, P2P finance, big data, blockchain, behavioral finance. Papers considered for Springer Digital Finance special issue.
Woman in Fintech Datathon
Organized by Women in Fintech Conference Committee with University of Tirana and University of Naples Federico II.
Participants: 18 teams, 46 participants (33% male) from Albania, China, Germany, Ghana, Greece, Italy, India, Kosovo, Lithuania, Romania. Task: Quantitative analysis of financial inclusion using World Bank Global Findex database.
digital-ai-finance.github.io/COST-Fintech-AI-in-Finance/datathon
Exploitation Platforms
Quantinar Platform
Data science education platform providing lecture-based learning with real-world project execution. Developed by Wolfgang Karl Hardle and team from Humboldt University Berlin. Course-lets span machine learning, fintech, cryptocurrencies, data science, blockchain, and statistics.
Quantlet Platform
Web-interface for freely exchanging numerical methods ("Quantlets"). Features: Open access to research/teaching codes, reproducibility of results, full GitHub integration, text mining pipeline, standardized code audits.
Meetup Community
Members: 2,080+ | Events: 150+ organized meetups
Interactive platform providing updates on talks, paper presentations, conference events, and working group meetings. Connects researchers and interested individuals on finance and AI topics.
Fintech.mk Community (North Macedonia)
First fintech community in North Macedonia, a non-profit organization providing IT and consulting services. 28 members, 13 articles published in English and Macedonian.
Data Juice Lab - XAI in Practice
Application of XAI knowledge from WG2 seminars in daily business practice. Vice CEO Marcin Chlebus applied XAI methods in FinTech projects covering Risk Modelling, Marketing campaigns, HR, and CRM.
Career Advancement & Promotions
Several researchers received promotions during the Action, demonstrating the network's significant impact on academic careers:
- Simon Trimborn - Promoted to Assistant Professor, Amsterdam School of Economics, University of Amsterdam (2022)
- Wolfgang Karl Hardle - Promoted to Visiting Scholar and YuShan Scholar, National Yang Ming Chiao Tung University (2022)
- Jorg Osterrieder - Promoted to Associate Professor (Finance/AI), University of Twente; also appointed Professor of Sustainable Business, Bern University of Applied Science
- Branka Hadji Misheva - Promoted to Associate Professor (Applied Data Science/Finance), Bern University of Applied Science
- Alla Petukhina - Promoted to Lecturer/Researcher, HTW Berlin (2021)
- Professor Tom Pilhal - Undergoing habilitation process to Associate Professor with COST Action support
Rapporteur Assessment of Career Impact
"The extent to which the Action has advanced the careers, skills and networks of researchers including ECIs is Very Good. This wider-range outcome demonstrates an effective and significant impact through its capacity to harness research and knowledge exchange to support academic career progression and promotion of a sizeable cohort of academic researchers."
Emerging Topics & Future Directions
Action-Identified Emerging Topics
- Increasing use of AI/ML in financial institutions for forecasting, portfolio management, fraud detection, customer service
- Blockchain technology in finance: digital currencies, smart contracts, DeFi
- Explainable AI (XAI) as regulatory and societal requirement
- Privacy and data protection in financial AI and fintech
Rapporteur's Additional Recommendations
"Driven from the exponential growth in Generative AI in the past 9-12 months, it is strongly suggested that the research directions are adjusted to clearly incorporate the impacts in areas such as:"
- Synthetic ID fraud - GenAI-enabled identity fabrication
- Privacy Enhancing Technologies (PETs)
- GenAI-resilience in KYC/CDD - Know Your Customer / Customer Due Diligence
- AI/ML impact on privacy and policy
Action Chair Response to Issues
Action Chair Statement
"We are pleased to report that our Action, COST Action CA19130 Fintech and Artificial Intelligence, has experienced significant growth and success. Our collective output and achievements have exceeded our initial objectives, and we do not currently see any issues that could potentially endanger our progress."
"In the grant period, we aim to complete the majority of our objectives in GP3, and we are excited to use GP4 to extend and improve upon our work, as well as to incorporate new challenges and research topics that have arisen as a result of our research efforts."
Rapporteur's Assessment of Plans
"The plan proposed is satisfactory, also given the sizeable progress recorded throughout the lifespan of the project."
Sample Publications (from 100 Peer-Reviewed)
The Action reported 100 peer-reviewed publications co-authored by at least two Action participants from two different countries. Selected highlights:
Liiv, I. - "Data Science Techniques for Cryptocurrency Blockchains" - Springer Singapore, Behaviormetrics Series. DOI
Jalan, A., Matkovskyy, R., Poti, V. - "Shall the winning last? A study of recent bubbles and persistence" - Finance Research Letters. DOI
Jalan, A., Matkovskyy, R., Urquhart, A. - "Demand elasticities of Bitcoin and Ethereum" - Economics Letters. DOI
Chen, Y., Bredin, D., Poti, V., Matkovskyy, R. - "COVID risk narratives: a computational linguistic approach to econometric identification of narrative risk during a pandemic" - Digital Finance. DOI
Spilak, B., Hardle, W.K. - "Tail-Risk Protection: Machine Learning Meets Modern Econometrics" - Encyclopedia of Finance, Springer. DOI
Ben Amor, S., Althof, M., Hardle, W.K. - "Financial Risk Meter for emerging markets" - Research in International Business and Finance. DOI
Li, W., Paraschiv, F., Sermpinis, G. - "A Data-driven Explainable Case-based Reasoning Approach for Financial Risk Detection" - arXiv. DOI
Hausler, K., Xia, H. - "Indices on cryptocurrencies: an evaluation" - Digital Finance. DOI
Wahlstrom, R.R., Paraschiv, F., Schurle, M. - "A Comparative Analysis of Parsimonious Yield Curve Models" - Computational Economics. DOI
Ren, R., Althof, M., Hardle, W.K. - "Financial Risk Meter for Cryptocurrencies and Tail Risk Network-Based Portfolio Construction" - Singapore Economic Review. DOI
Overall Conclusions
Rapporteur's Final Assessment
"Overall, the COST FinTech and AI in Finance Action has so far made impressive headways in mobilising a robust FinTech community and progressing the State of the Art in FinTech-related data science and successfully disseminating a large body of associated and highly consequential research."