⭐ Excellence: Research Programme & Methodology

13 sections from the Description of Action (Part B)

1.1.1. Introduction, objectives and overview of the research programme

Introduction and Timeliness. A competitive European financial sector is vital for the modernisation of the

European economy across sectors and to turn Europe into a global digital player. The term Digital Finance refers

to the rapid development of new technology, goods, and business models that have taken place in recent years.

We have identified the five most pertinent areas within this domain:

● Towards a European financial data space.

● Artificial intelligence for financial markets.

● Towards explainable and fair AI-generated decisions.

● Driving digital innovations with Blockchain applications.

● Sustainability of Digital Finance.

What they have in common:

● They are all key strategic priorities of the European Commission over the next five years.3

● They contribute to the UN Sustainable Development Goals.4

● Europe must invest significantly in them over the next five years if it is to remain globally competitive

● They are characterised by a significant shortage of skilled labour.

● Initial progress has been made in academia, but there are still numerous unanswered research questions.

● They have the potential to revolutionise the Finance industry with new technologies, business models,

and products, while strengthening the resilience of Europe.

● They are the foundation for a new generation of PhD candidates and training in Digital Finance.

Considering these developments across industries and within the financial sector, it is absolutely essential to work on

those research topics now and to train new PhD graduates, because:

● Digital Finance has already changed the way the Finance industry works.

● To deal with the realities of academia and industry, PhD graduates in Finance will be required to acquire

the skill set of Digital Finance.

● There is a substantial research gap in academia that needs to be resolved now by academics and a new

generation of Digital Finance PhDs to keep Europe's Finance industry competitive.

Network. For this purpose, we have gathered an internationally recognized network consisting of nine leading

European Universities (University of Twente, WU Vienna, , University of Naples, Kaunas University of Technology,

Bucharest University of Economic Studies, Babes-Bolyai University ,Bern Business School, Poznan University of

Economics and Business and University of Kaiserslautern-Landau), all ranked among the top 200 universities

globally in their fields, three major international corporations (Swedbank, Raiffeisen Bank and Deutsche Bank)

with a significant R&D presence across Europe, two SMEs (Cardo AI and Royalton Partners) being some of the

most innovative ones in their field, three large and internationally renowned research centres (ARC Greece, EIT

Digital and Fraunhofer Institute), the Bank for International Settlements and the European Central Bank, as one of

the seven principal decision-making bodies of the European Union and the Euratom as well as one of the world's

most important central banks. The results of the research carried out within DIGITAL are of substantial interest to

three leading European-wide research networks that our members either lead or serve on the management

committee for: COST Action CA19130 Fintech and AI in Finance (240 researchers across 39 European countries),

European Consortium of Mathematics for Industry (200 researchers across Europe) and the European Consortium

of Innovative Universities (13 European Universities). It is only through a network that incorporates the

expertise of all distinct shapers of the financial industry (technology experts, academics, Fintechs, domain experts,

incumbents, regulators, civil society) that we can see a comprehensive shift towards innovation and digitalization

of a sector that is notoriously averse to change.

3 A European Approach to artificial intelligence ( https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence), the EU Digital

Finance Package (https://finance.ec.europa.eu/publications/digital-finance-package_en) 4 https://sdgs.un.org/goals

4 https://sdgs.un.org/goals

8

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

Objectives. Today, Digital Finance does not exist as a standalone research discipline, despite many research gaps,

the EU’s key strategic priorities and the urgent needs from industry. DIGITAL will overcome this and

significantly advance the methodologies and business models for Digital Finance through the use of five

interconnected and coherent research objectives and a total of seventeen Early Stage Researchers (DCs) hired by the

beneficiaries, both from academia and industry. The main objectives are:

● Towards a European financial data space. Ensure sufficient data quality to contribute to the EU's efforts

of building a single digital market for data (WP 1).

● Artificial intelligence for financial markets. Address deployment issues of complex artificial intelligence

models for real-world financial problems (WP 2).

● Towards explainable and fair AI-generated decisions. Validate the utility of state-of-the-art explainable

artificial intelligence (XAI) algorithms to financial applications and extend existing frameworks (WP 3).

● Driving digital innovations with Blockchain applications. Design risk management tools concerning the

applications of the Blockchain technology in Finance (WP 4).

● Sustainability of Digital Finance. Simulate financial markets and evaluate products with a sustainability

component (WP 5).

Research Training for Digital Finance. The network will specifically train young researchers in R&D topics that

cover the multiple disciplines required in the rapidly evolving field of Digital Finance substantially going beyond

the traditional Finance PhD education in a wide range of inter-sectoral applications: data quality, Artificial

Intelligence (AI) and Machine Learning (ML), Explainability of AI (XAI), Blockchain applications and sustainable

finance; all of which are required for a wide range of industrial (financial products, risk management, customercentric

products, enhanced processes, and improved services) and scientific (new AI techniques, new business models, and

enhanced modelling) applications, necessitating new scientific insight, new training courses, and future specialists in

the field.

Need for an Industrial Doctoral Network. The European Finance industry needs to compete on a global scale. To

overcome key hurdles which financial service companies will face in the near future, they will have to find answers to

(WEF 20205):

● Data quality issues related with the increasing dimensionality of financial data.

● Deployment issues of complex models in real-world applications.

● Deficits in trust and user adoption of AI-supported financial products.

● Potential data or algorithmic bias inherent in AI models.

● Labour shortage: AI leaders overwhelmingly argue that access to talent represents a key obstacle to the

digitization efforts in finance, as more sophisticated solutions demand different employee capabilities.

All of those hurdles towards scientific, societal and economic/ technological impact will be solved in DIGITAL.

Research programme. The strategic priorities of the European Union, industrial needs and academic research

gaps lead to the scope of our research programme. As detailed in Part A, within the structure of DIGITAL's work

packages, four early stage researchers (DCs 6, 8, 13, 15) will focus on proposing novel methodologies and

applications to address the key data quality issues associated with high-dimensional, high-variety, and

highvelocity datasets. In addition, deployment issues with cutting-edge machine learning, deep learning, and

reinforcement learning methods must be addressed in order for methodological advances to be useful in real-world

applications. DCs 12, 14 will be solely devoted to addressing the most significant obstacles to industry adoption

of technological innovation. DCs 1, 9, 16, 17 will conduct work closely related to this. They will establish the

principles of a trustworthy and secure AI for financial applications and define the methodologies that can satisfy

the explainability requirements of various stakeholders within the financial sector. In this context, the outputs will

become central to the applicable regulation of this technology. DCs 3, 5, 7 will examine the efficacy and impact of

Blockchain adoption in financial markets and propose risk management tools to address some of the main concerns

around this technology (e.g., financial stability and fraud detection). Additionally, guidelines for a supervisory

approach to machine learning for digital finance, and various perspectives will be explored. . DCs 2, 4, 10, 11 will

focus on sustainable financial products, simulating markets and designing recommender systems to analyse the

effects of adopting principles of sustainable finance philosophy.

This collaborative effort will serve as the foundation for a new qualification standard that will provide European

firms the competitive advantage they need to compete in the area of Digital Finance. To achieve DIGITAL's research

5 WEF 2020, https://www3.weforum.org/docs/WEF_AI_in_Financial_Services_Survey.pdf

9

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

and training objectives, all projects must be handled collectively for an integrated view of Digital Finance;

standalone projects will not have a significant impact.

The 17 Individual Research Projects (IRPs) are grouped into five inter-connected WPs. WP 1 lays the foundation

for all other WPs. Data is in the centre of the digital transformation (A European strategy for data, European

Commission 20206 ), though utilising it is accompanied by many challenges. By ensuring that all dimensions

(accuracy, consistency, completeness, currency, volatility and timeliness) of data quality are satisfied, that we have

sufficiently large, high-quality data available and that we can detect dependencies in high dimensions, high frequency

and high veracity financial data, DIGITAL is in a position to move towards a European financial data space and

make use of substantially more data sources than today.

WP 2 ensures the deployability of AI models. AI has improved decision making in numerous areas, including risk

management, compliance (anti-money laundering, fraud prevention, KYC), trading strategies and personalised

banking and advice (A European approach to AI, European Commission 20217), and yet, adoption of AI-based

tools in practice has still been rather slow. With DIGITAL, we will describe the primary challenges and

opportunities for industry's adoption of technological development, encourage a larger deployment of stateof-

the-art ML models in real-world financial applications and simulate the market environment in Reinforcement

Learning (RL) applications for market applications, thereby removing the primary barrier to the application of this

technology in finance.

WP 3 verifies that deployed AI models are explainable, trustworthy and unbiased. AI-driven innovation can bring

enormous benefits but such complex solutions are often referred to as “black boxes” because typically it is difficult

to trace the steps the algorithms took to arrive at its conclusions (The European Approach to Excellence and Trust,

European Commission 20208). DIGITAL will describe how well XAI tools meet the explainability requirements of

various financial value chain stakeholders, develop non-perturbation-based XAI methods that preserve the natural

time ordering and dependence structures of the data and create methodologies to ensure that algorithmic systems do

not produce socially biassed outcomes that exacerbate inequalities.

WP 4 uses the Blockchain technology for digital innovations. Blockchain is another driver of the technological

change of the financial ecosystem (The European Commission’s Blockchain strategy, 20219), though there are no

systematic studies that assess whether the benefits outweigh the costs. DIGITAL will analyse the efficiency of

financial service providers that adopt the blockchain technology, contribute to more robust and efficient financial

markets by understanding how to tokenize financial assets, reduce the risk of fraud and highly volatile crypto

assets and establish guidelines for a supervisory approach to machine learning for digital finance. WP 5 makes

Digital Finance more sustainable. Climate change and environmental degradation are becoming an existential threat

to Europe and the world (European Green Deal, 201910), though inevitably, transitioning to a sustainable future with

inclusive, green economies and resilient ecosystems is associated with many challenges. Within DIGITAL, we will

make financial strategies for sustainable investing more objective, optimised, integrated, and operational, measure

the social and environmental effects of sustainable finance, define the primary characteristics of sustainable

financial assets and instruments and overcome the most significant obstacles to implementing sustainable finance

strategies.

WP 6 will provide Doctoral Training in a new, innovative and interdisciplinary training structure for Digital Finance

scholars. The required subjects are currently not available or are so far included in very different academic

disciplines. They will be combined in a coherent study and training plan that will result in the formation of highly

qualified professionals as well as a long-term sustainable PhD programme, thus guaranteeing the competitiveness

of the European industry in a strategic domain such as that of Digital Finance, following the MSCA guidelines on

supervision.11

WP 7 ensures communication, dissemination and exploitation of our research results, both in the scientific and

public domain with substantial impact due to our focus on industry prototypes and use-cases.

DIGITAL’s research and training framework is tightly integrated into the European Digital Finance Package (EC

202012), with substantial involvement of the European Central Bank as one of our partners and will contribute

to one of the four EU’s key strategic orientations,13 namely the promotion of an open strategic autonomy through

6 https://digital-strategy.ec.europa.eu/en/policies/strategy-data

7 https://digital-strategy.ec.europa.eu/en/policies/european-approach-artificial-intelligence

8 https://ec.europa.eu/info/sites/default/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf

9 https://digital-strategy.ec.europa.eu/en/policies/blockchain-strategy

10 https://ec.europa.eu/info/strategy/priorities-2019-2024/european-green-deal_en

11 https://op.europa.eu/en/publication-detail/-/publication/bb02d56e-9b3c-11eb-b85c-01aa75ed71a1/language-en

12 https://finance.ec.europa.eu/publications/digital-finance-package_en-

13 https://www.eeas.europa.eu/sites/default/files/horizon_europe_strategic_plan_2021-2024.pdf

10

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

the development of key digital, enabling and emerging technologies, sectors and value chains that further accelerate

the digital and green transition of Europe.

1.1.2. Pertinence and innovative aspects of the research programme

State-of-the-art research and beyond. Both academia and industry are increasingly faced with data characterized

with staggeringly high number of dimensions, high variability and high veracity. Utilising such data is accompanied

by many challenges. In WP 1, an unprecedented collection of innovative data quality methodologies and data

augmentation techniques will be applied to a wide variety of industry-relevant datasets. Overcoming the

obstacles of data quality and availability (through novel or extended NLP methods,14 deep generation of data,15 and

anomaly- and dependence detection models baked on network concepts) will contribute to the literature and further

provide a valuable tool for the European Finance industry to enhance product offerings, reduce financial market

risks, and work toward a European financial data space.

In WP 2, the emphasis will be placed on the deployment of complex AI models to pertinent financial problems.

For financial applications such as risk management, trading strategies, and client-centric financial products, AI

models trained and tested in closed, academic settings have shown great promise. Yet, real-world applications (in

open environments) are more challenging. Using industry-ready use cases, we will demonstrate the viability of

novel dynamic rating models,16 automated trading platforms,17 and market environments for RL agents18 in

real-industry settings for the first time. WP 2 will provide first-of-its-kind qualitative analysis of the primary

obstacles to deploying innovative technologies in industry and propose new solutions for resolving these obstacles

which in turn is a crucial step towards a widespread adoption of complex models in the financial sector.

WP 3 will address the crucial question of how to build trust in human-centric AI models as opposed to the currently

widespread AI black boxes,19 which do not meet the modern European requirements of explainability, trust and

unbiasedness. This WP will also utilise the outcomes of WP 2, as the precise methodologies for explainable AI will

depend heavily on the specific models developed. We will validate the applicability of state-of-the-art XAI algorithms

to financial applications and extend XAI frameworks, ensuring that complex models applied to financial use cases

satisfy the explainability requirements of different stakeholders within the finance value chain and do not reinforce

social biases. A qualitative evaluation of the comprehensive frameworks' insights into explainability will be made in

comparison to the baseline models. Through industry-ready use cases, we will demonstrate for the first time ever

the viability of the proposed framework for audience-dependent explanations,20 the novel timeseries XAI

methods, and the fair algorithmic designs.21

In addition, WP 4 will utilise the research progress made in the first three WPs to make use of Blockchain technology

and to design systematic studies aimed at advancing financial digital innovation. These studies will serve as a crucial

input for a new infrastructure that will be utilised in a wide variety of IT domains. This infrastructure will account for the

integration of multiple data sources, define a standard dictionary, eliminate ambiguity, and enable other teams to access

all customer data from a centralised repository, thereby ensuring interoperability. Defining and monitoring efficiency

measures will ensure the quality of the proposed frameworks. In addition, a qualitative evaluation will be conducted

based on the comparison of our proposed frameworks that go beyond the state of the art of traditional standard

approaches.22 This is intended to serve as the basis for a new blockchain-based financial infrastructure and as a

European industry standard for blockchain applications. In addition, WP 4 will also focus on proposing novel risk

management solutions to some of the main concerns around blockchain applications in finance, like fraud detection and

financial stability. Utilising concepts from network theory and a set of exploratory tools to improve statistical models, we

will develop industry-ready use cases for fraud detection in financial networks23 and propose a comprehensive and

dynamic risk index for cryptocurrencies.24

In WP 5, a novel study will be conducted to simulate and evaluate markets in order to replicate the relationship

between banks, financial institutions, and their retail and business clients in a sustainable environment that takes into

14 Li et al. (2022). Incorporating Transformers and Attention Networks for Stock Movement Prediction, Complexity. https://doi.org/10.1155/2022/7739087

15 Wiese, M. et al. (2020). Quant GANs: Deep generation of financial time series. Quantitative finance. 20. https://doi.org/10.1080/14697688.2020.1730426

16 Vana et al. (2021). Dynamic modelling of corporate credit ratings and defaults. Statistical Modelling. https://doi.org/10.1177/1471082X211057610

17 Cohen, G. (2022). Algorithmic Trading and Financial Forecasting Using Advanced AI Methodologies. https://doi.org/10.3390/math10183302

18 Karpe et al. (2020). Multi-agent reinforcement learning in a realistic limit order book market simulation. ICAIF '20: Proceedings of the First ACM International

Conference on AI in Finance. 30. https://doi.org/10.1145/3383455.3422570

19 Rudin, C. et al. (2019). Why are we using black box models in AI when we don't need to? Harvard Data Science. https://doi.org/10.1162/99608f92.5a8a3a3d

20 Hadji Misheva, B. et al. (2021). Audience-Dependent Explanations for AI-Based RM Tools: A Survey. Frontiers. https://doi.org/10.3389/frai.2021.794996

21 Hajian, S. et al. 2016. Algorithmic Bias: From Discrimination Discovery to Fairness-aware Data Mining. 22nd ACM SIGKDD 2016. Association for Computing

Machinery, New York, NY, USA, 2125–2126. https://doi.org/10.1145/2939672.2945386

22 Bruno Biais et al, The Review of Financial Studies, Volume 32, Issue 5, May 2019, Pages 1662–1715, https://doi.org/10.1093/rfs/hhy095

23 Ashfaq, T. et al. (2022). A Machine Learning and Blockchain Based Efficient Fraud Detection Mechanism. MDPI. 22. https://doi.org/10.3390/s22197162

24 Trimborn, S. and Haerdle, W. (2018). CRIX an Index for Cryptocurrencies. Journal of Empirical Finance. 49. https://doi.org/10.1016/j.jempfin.2018.08.004

11

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

account various factors (e.g., Green AI25 and green credit score26). To evaluate markets, numerous indicators and

metrics, including loan interest rate, customer attrition, CO emissions, customers' access to credit, and firms' profits,

2

will be compared. We will use agent-based modelling for the first time to simulate different market scenarios in which

agents involved in the industry take sustainable actions.27 Long-term financial growth will be analysed, and the

findings will inform the development and modification of industry policies and strategies. This will ensure that

sustainability considerations are integral to the Digital Finance industry, contributing to the European Green

Deal in the process.

Existing doctoral training programmes. Across Europe and globally, many separate PhD graduate schools exist in

the different disciplines involved, most notably Finance (incl. many Ivy league Finance programmes), Computer

Science, Applied Mathematics and Economics. Those are very-well established programmes with a long history

of achievements. Two past doctoral training programs related to Finance and Data Science topics have been funded

by the Marie Skłodowska-Curie Innovative Training Network and have offered interdisciplinary training to

candidates (Data Engineering for Data Science (DEDS),28 Training for Big Data in Financial Research and Risk

Management (BigDataFinance)29).

Gaps in existing training schemes. Existing programs, however, are insufficient for the new era of Digital

Finance, which requires a comprehensive set of intersectoral and interdisciplinary competencies. None of them can

provide the necessary skill set, as the necessary courses are either not provided or taught in separate silos, not to the

same cohort of DCs. Courses that are essential, but not offered in classical Finance PhD programs are e.g.: big data

analytics and data management, deep data generation, reinforcement learning, XAI, blockchain technologies and

network theory. These gaps will be closed by our doctoral training, which combines seven disciplines with a unique

training offer, also for transferable skills, from some of the leading European institutions, such as full access to all

training courses offered by the European Central Bank, EIT Digital, the European Consortium for

Mathematics for Industry, and world-renowned universities.

We fill the interdisciplinarity void with a network of seven disciplines, Economics, Finance, Management,

Economic Development, Computer Science and Informatics, Applied Mathematics, and Political Science. In terms

of substantial inter- and intra-sectoral training (European Central Bank, Bank for International Settlements,

DeutscheBank, ARC, EIT Digital, Fraunhofer Institute), as well as training for sustainability and entrepreneurship,

we will close the gap that exists in current programs. Finally, our unique collaboration with industry ensures both

training in transferable skills and academic training related to the industry, most notably the European Central

Bank, making their in-house courses available.

Long-term sustainability of the Doctoral Training. All partners have fully committed to the work of the University

Business Forum30 and the results of the EUA DOC-CAREERS31 project, and will support the development of the

university partners' curricula and doctoral training with a long-term sustainable doctoral course offering and ensure

that the acquired skills more closely match the needs of our European industry. All courses and training offers will

last beyond the life-time of the MSCA Doctoral network and doctoral training rules will be harmonised, as

already agreed upon by all partners.

An ambitious project. DIGITAL is ambitious in that it seeks to train doctoral candidates and develop and disseminate

research on all topics pertaining to Digital Finance. The ideal trained personnel will be capable of addressing data

quality issues, developing, implementing, and supervising AI models, designing, proposing, and critiquing XAI tools,

implementing and managing Blockchain applications, and facilitating the transition of the industry's current tasks

toward the adoption of sustainable finance.

25 Schwartz et al. (2020). Green AI. Communications of the ACM, 63. https://doi.org/10.48550/arXiv.1907.10597

26 Akomea et al. (2022). A review of studies on green finance of banks, research gaps and future directions. J of Sustainable Finance & Investment, 12.

https://doi.org/10.1080/20430795.2020.1870202

27 Abdou, M., et al (2012). Designing and building an agent-based model. In Agent-based models of geographical systems (pp. 141-165). Springer, Dordrecht.

28 https://cordis.europa.eu/project/id/675044

29 https://cordis.europa.eu/project/id/955895

30 http://ec.europa.eu/education/higher-education/doc1261_en.htm

31 http://www.eua.be/eua-work-and-policy-area/research-and-innovation/doctoral-education/doc-careers

12

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

1.2.1. Overall methodology

The network follows an approach of broad comprehensive training complemented by cutting-edge research projects

executed by 17 DCs. Researchers will not only develop key scientific skills and propose novel technologies that are

on the frontiers of big data-, AI-, blockchain- and sustainability research, but they will also build easily deployable

solutions to the key hurdles that finance service providers face in using these technologies in production,

disseminating this to all stakeholders, incorporating feedback from industry and policy makers.

The central research question in DIGITAL is how innovative technologies, like big data, artificial intelligence and

blockchain, can be used to support Digital Finance in view of the emerging complexities: (i) high-dimensional,

highvariety, high-velocity dataset; (ii) limited samples of high-quality data to train various ML models, (iii) no

comprehensive pipeline for building and deploying complex ML models in real settings, (iv) no explainability

techniques that are specifically tailored to financial datasets and satisfy the explainability needs of various financial

stakeholders, (v) no industry standard for blockchain applications and (vi) no common ESG scoring framework.

All of these complexities are methodological in nature and in order to tackle them, DIGITAL will carry out pragmatic,

data-focused, inductive research using a combination of research strategies (case studies, experiments and

actions) and methods enriched with continuous cooperation with and feedback loop from industry and

regulators. We elaborate on the central layer of the research methodology, which combines academia, industry,

dissemination, and training with seven building blocks.

5. Dissemination and Communication methods. Our methodology relies on substantial dissemination and

communication efforts (such as more than 120 academic events organised by our extended network since 2020),

which will enable us to collect feedback, stay abreast of the most recent developments, and shape the future of this

research direction. DCs receive extensive feedback from academics and other stakeholders, allowing them to expand

or shift their research focus and methods, as well as validate the effectiveness of their approaches.

6. Integration into doctoral training. Our methods also anticipate feeding back into doctoral education. All

DCs will routinely present their findings at DIGITAL seminars, and supervisors will be expected to incorporate new

findings into doctoral training programs. This will be accomplished primarily through the new modules developed

by the network (see Table 1.3.c), which will be continuously updated to include the new methods, technologies,

solutions, tools, and business models created by the network.

7. Cooperation with leading institutions. Our methodology is also strongly based on working together with

very large networks of researchers (more than 400 academics in COST, 170 institutions in ECMI) to collect

feedback and learn the latest techniques, as well as leading institutions, such as the European Central Bank and the

Bank for International Settlements.

Challenges. There are both data availability and methodological challenges. Data has already been used, but if

some new data cannot be shared easily, we will provide specialised legal solutions (NDAs), and acquire the required

licences. Additionally, in case the problem remains unsolved, we will employ data anonymization techniques that

enable data sharing or resort to synthetic data generation techniques (GANs). On the methodology-related

challenges, in cases in which the planned techniques do not lead to the envisioned outcomes, we will expand the

scope of models to be trained and tested and reevaluate the planned effort. In case of any deployment issues (i.e.

the developed methodologies are not fully suited to the IT constraints in which finance service providers operate),

we will offer the solution as a standalone product rather than as an incorporated service.

1.2.2. Integration of methods and disciplines to pursue the objectives

Digital Finance necessitates the integration of methods, information, data, techniques, tools, perspectives,

concepts, and theories from the seven essential disciplines that we combine in our network. Those are Economics,

Finance, Management, Economic Development, Computer Science and Informatics, Applied Mathematics, and

Political Science. This emphasis on interdisciplinarity will be incorporated not only into the design and

implementation of the research framework, but also into the training program.

Interdisciplinarity of Research Methods.

● Each IRP is developed at the intersection of multiple disciplines and requires input and feedback from

academic experts conducting research in various fields. Thus, each DC will be assigned three supervisors

with complementary skills and experience to receive inter-disciplinary guidance to meet research objectives.

In addition, all DCs will have access to our three international partner networks, with over three hundred

researchers from around the world working on AI and fintech technologies.

● All IRPs have an applied focus, each addressing a specific obstacle companies face when deploying

innovative financial technologies. To provide each researcher with the necessary applied guidance, an

industry supervisor will be assigned to lead and oversee the CDP's applied objectives.

● Regulatory and supervisory bodies play a crucial role in defining the business models of financial service

providers. To account for the regulatory/supervisory discipline as well, researchers will have the opportunity

to validate their research through secondments at the European Central Bank.

Interdisciplinarity of Training.

● All DCs will take an interdisciplinary set of courses, such as: classical finance (risk management, trading

and portfolio optimisation, etc.), computer and data science (synthetic data generation, blockchain in digital

14

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

finance, RL for finance, etc.), statistics and econometrics (risk quantification, dependence structure in

highfrequency data, etc.), database management systems (big data, etc.), and AI (digital finance regulation,

monitoring practices, etc.).

● All DCs will receive transferable skills training, including tutorials from industry. This training will also

be highly interdisciplinary, covering very diverse industry areas, such as large banks, innovative Fintech

companies, consulting companies, international research centres and the European Central Bank.

Industry will focus on providing data-sets and applied research topics, as well as converting prototypes and

use cases to financial products, and training DCs in transferable skills and applied research.

Table 1.2.a Role of each consortium member in the research program

Consortium member Role

All partners Co-setting the specific research objectives within their respective IRPs, co-supervising DCs, and collaborating with

their respective industry and academic co-supervisors.

Project coordinator. WP 6, 8 Lead. Contribute data on crowdfunding, financial time series and ESG scores as well as

theoretical and applied expertise in ML modelling, Fintech business models, sustainable finance, blockchain

University of Twente

technologies.

WP 2 Lead. Contribute theoretical and applied expertise in building supervised and unsupervised ML systems, AI and

WU Vienna

financial technology, applied mathematical programming.

WP 5 Lead. Contribute theoretical and applied expertise in analysing complex categorical data structures, survey

University of Naples

analytics, latent models, link functions and gender studies.

Kaunas University of Ensure coherence in the research and applied objectives within each IRP. Contribute theoretical and applied expertise

Technology in blockchain-based conditional payments, stochastic modelling, supervised learning algorithms.

Bucharest University WP 4 Lead. Contribute data on cryptos, SMEs financial data as well as theoretical and applied expertise in building

of Economic Studies predictive ML and DL models for finance, clustering and community detection algorithms, case-study analysis.

Babes-Bolyai WP 1 Lead. Contribute theoretical and applied expertise in spatial econometrics, text mining and NLP, fraud detection

University systems, panel data modelling.

Poznan University Contribute data on cryptos, theoretical and applied expertise in statistics, cryptocurrencies, credit risk, predictive

analytics in digital finance

University of Contribute with theoretical and applied expertise in machine learning, time-series forecasting, credit portfolio analyses,

Kaiserslautern-Landau data analytics, impact of innovation in Fintech

Cardo AI Contribute data on private debt investment and expertise in building software for transparent data analytics.

Raiffeisen Bank Contribute data, training on financial research, and overall expertise in retail banking, asset management.

Swedbank Contribute data, training on retail banking, and expertise in asset management and other financial services.

DeutscheBank Contribute with various datasets, provide expertise on applications of AI, especially anomaly detection and early

warning systems, as well as expertise on predictive analytics, semantic analysis and risk management.

Royalton Partners Contribute data on cryptos and expertise in portfolio and risk management, compliance and regulatory reporting.

Bern Business School WP 3, 7 Lead. Contribute data on SME and personal loan performance and ESG scoring as well as expertise in

algorithmic design, ML and DL models for credit risk management, deploying explainable AI systems.

ARC Contribute data and overall expertise in applied research, product development and infrastructure service provision

for digitising processes and societies.

EIT Digital Co-setting the specific training objectives within WP 7; ensure the excellence of the training programme, by bringing

in their experience from the European Master in Digital Finance.

Fraunhofer Institute Contribute data and overall expertise in enabling efficient and sustainable transfer of scientific knowledge into

commercial use. Exposure to a world-leading applied research environment

European Central Contribute various macro-economic datasets as well as overall expertise in banking supervision, statistics, foreign

Bank exchange operations, payment systems and defining monetary and fiscal policy.

Bank for International Contribute macro-economic datasets, ongoing projects as well as overall expertise in banking supervision, statistics,

Settlements foreign exchange operations, remittances, payment systems, central bank digital currencies and defining monetary and

fiscal policy.

1.2.3. Gender dimension and other diversity aspects

In addition to all WPs strictly adhering to diversity considerations, WP 3 is specifically devoted to fair, trustworthy,

and unbiased artificial intelligence.

Gender dimension and other diversity aspects in the project's research and innovation activities. DIGITAL will

represent EU values of sex and gender equality and diversity, believing that research and innovation must recognize

15

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

and integrate these issues for societal relevance and quality, following the EU policy report on Gendered

Innovations.32 Sex and gender will be incorporated into its research design. DIGITAL's strategy will be:33

1.2.4. Open science practices

Adherence to all mandatory and recommended open science practices. DIGITAL has extensive experience with

and will adhere to all mandatory open science practices outlined in the Horizon Europe Programme Guide36, the

obligations in Articles 17 of the Model Grant Agreement, and the corresponding guidelines, monitored by the SB.

We will even go above and beyond these requirements by extending the adherence to all associated partners.

All partners will also adopt all recommended practices and attend a mandatory open science course.

Mandatory training and institutional changes. We will actively change all beneficiaries' institutional governance

to ensure open science practices last beyond the project's funding.

Open Science practices are a common standard in our discipline. Early and open sharing of research will be done

via publishing preprints at arxiv and pre-registering research plans and reports at OSF and Protocol Exchange.

Research Data Management is achieved by using DOIs for all digital objects and metadata frameworks (incl. data,

publications and other research outputs), using free and open source licences. Zenodo is our trusted repository.

Based on the European Open Science Cloud (EOSC), reproducibility will be guaranteed, via using the DMP,

FAIR principles, and GitHub for storing and managing code. Open access to research outputs will be mandatory

for all partners, via publishing in open-access and open-peer review journals such as Open Research Europe

(Consortium members already have a leading role).

Citizen, civil society and end-user engagement will be ensured via co-designing, co-creating and co-assessing

activities. We have dedicated workshops with civil society as well as an annual Open Science day. Already in 2010,

our partners (coordinated by ASE) have built their own open science ecosystem for sharing datasets and software

with more than 1000 software pieces and 2000 datasets collected, used by more than 500 researchers from our

disciplines.37

1.2.5. Research data management and management of other research outputs

DIGITAL's data management planning (WP 8) relies on the infrastructure, tools, and support of each partner and

their data management departments (incl. dedicated Data Stewards). During months 1-6, data for each IRP will be

collected and integrated into DIGITAL's systems. Tests will ensure data is securely stored and shared. DIGITAL's

main goal in WP 1 is to propose a FAIR dataset (including software and metadata) for each IRP, accounting for:

5. Data curation/Cost of data curation. DIGITAL will recommend DMP storage and backup best practices,

which beneficiaries and partners will implement under the supervision of computer engineers and data management

experts. Partners will provide the use of resources at no cost. The SB will monitor data curation quality and DMP

compliance. Data will be permanently archived in OpenAIRE to support European research even after the project. A

detailed DMP inline with GDPR will be written that follows open access principles to research data, being updated

regularly according to new types and volumes of data used in each IRP. The DMP will contain sufficient

documentation of data and metadata, data workflows, software licensing, privacy policies, and intellectual property

rights. It explains when commercial data can be reused and how the data meets FAIR and EOSC standards.38

1.2.6. Artificial Intelligence

DIGITAL proposes developing and employing a wide range of AI systems (WP 1 - 5) for use in Digital Finance,

while adhering to the European Commission's AI39 guidelines on a consistent basis. By design, technical robustness

will be ensured (WP 3). These systems will be developed to become:

5. Monitoring best practices on dissemination of IRP results; and 6. Reporting to the SB on common issues and

proposing an agenda for their meeting.

55 DIGITAL will comply to the gender balance while forming all management teams, following the European Commission guidelines (https://research-

andinnovation.ec.europa.eu/strategy/strategy-2020-2024/democracy-and-rights/gender-equality-research-and-innovation_en ) 57 https://www.desca-

agreement.eu/desca-model-consortium-agreement/

35

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

Supervisory Board (SB). The Supervisory Board/ Steering Committee is composed of one representative from

each partner, all supervisors, all WP leaders, and two doctoral candidates who rotate annually. The SB is co-chaired

by the project coordinator and the EU representative. It will elect one representative to serve as the SB's diversity

coordinator. The SB will have a kick-off meeting at the start of the network activity, with bi-annual consecutive

meetings. The SB is in charge of approving and supervising the financial, marketing, and communication strategies.

In addition, monitoring and evaluating the progress of training and research milestones. The SB is also responsible

for serving as an internal audit team for the EB when it decides on budget allocation, organisational structure, and

consortium strategy.

External Advisory Board (EAB): The EAB will review DIGITAL's progress in training and research every six

months. The EAB will be composed of prominent academics and senior executives with dual responsibilities. They

will issue recommendations and brainstorm ideas for training and research programs, and they will also propose

corrective measures.

Doctoral Candidates Committee (DCC): The DCC has two members and one doctoral candidate representative

from each beneficiary, and its chair is chosen annually by the DCC's members. The DCC is responsible for

supervising training and facilities issues, ensuring DIGITAL's policies are consistent with their implementation,

engaging in discussions and voicing ideas and concerns to their representative on the SB, and ensuring engagement

of the DIGITAL network by promoting social events for doctoral candidates and all consortium members.

Research and Training Committee (RTC): The RTC is composed of two members each from the academic and

non-academic sector, and one representative from an associated partner. It is chaired by a designated training advisor

and co-chaired by a representative of an industrial partner. The RTC is responsible for setting the strategy on

organising courses and training activities, creating an individual training plan for doctoral candidates that takes into

account their career development, and assisting the consortium with supervision-related issues and conflicts.

Communication and Dissemination Board (CDB): The CDB will be chaired by a representative of an associated

partner and composed of a representative of DIGITAL's beneficiaries, a member representing the industrial partners,

another member representing the associated partners, and a member representing the doctoral candidates. The CDB

will be chaired by a representative of the beneficiaries, and its primary responsibilities will include ensuring that

multiple and appropriate audiences are targeted throughout the various communication campaigns, monitoring the

implementation of communication and dissemination strategies, and monitoring engagement from different

audiences in order to maximise the impact of DIGITAL results. A media communication officer will be appointed.

IP and Exploitation Team (ET): The ET will consist of a representative from each industrial partner and two

academic advisors designated by the partners. ET will be chaired by a representative of an industrial partner and will

be responsible for identifying potential products and services from each IRP's outputs, proposing and monitoring the

creation of roadmaps, prototypes, and software. In addition, through the adoption of open science practices, ET will

ensure the sharing of knowledge, skills, and data. All partners have substantial exploitation teams at their institutes.

To support us, we will have one designated IP officer and two officers for exploitation from within our network.

Project Coordinator Team (PCT): The PCT is chaired by the project coordinator. A dedicated technical manager

and financial manager will be assigned. The project coordinator will establish the network, manage the relationships

with the external advisory board and coordinate the overall work of the project and its implementation, both at the

technical and at the financial level. The coordinator will monitor the content and progress of each work package as

well as coordinate the cooperation among work packages and assure the effectiveness of the project, according to

appropriate measurement standards. In this respect, the coordinator will assign a technical manager to the project

which will work in coordination with the project coordinator to ensure that milestones are respected, and deliverables

are submitted in time with high quality.

Specifically, the technical manager will check with each partner ahead of the planned milestones and deliverables

whether they are going to be completed in time and with the required quality. The coordinator will monitor how the

financial resources allocated to each partner are employed, to ensure effectiveness of the project, keeping high quality

standards for the produced deliverables. In this respect, the coordinator will assign a financial manager to the project

which will work in coordination with the administrative offices of University of Twente to ensure financial

compliance and transparency. Specifically, the financial manager will check with the technical manager in advance

36

Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022

whether each partner respects the milestones and deliverables and in which proportion; and will assign the

corresponding budget.

The coordinator will be responsible for maintaining internal communications within the consortium, with the

supporting external parties and for reporting to the EC.

Figure 3.1. DIGITAL Network Organisation