4 sections from the Description of Action (Part B)
DIGITAL will produce a substantial amount of new doctoral training materials that will be taught at network-wide
events, made available to all. All partners make their internal and local doctoral training available to all DCs; the
IRPs and substantial secondments at world-class institutions like the ECB will provide substantial training.
Pillar 1: Training through research and mandatory scientific training. Each recruited researcher will develop
new scientific knowledge and skills through conducting original research under their IRPs. Researchers will need to
follow and pass five mandatory, foundation courses (local training, Table 1.3.a). DCs will have mandatory network
training events (Table 1.3.b). All network training events will feature presentations from DCs.
38 https://en.wikipedia.org/wiki/FAIR_data, https://eosc-portal.eu/, https://zenodo.org/, https://datacite.org/, https://www.dublincore.org/,
https://www.w3.org/2004/02/skos/, https://www.openaire.eu/about, https://www.budapestopenaccessinitiative.org/read/,
https://openaccess.mpg.de/Berlin-Declaration
39 https://eur-lex.europa.eu/legal-content/EN/TXT/PDF/?uri=CELEX:52018DC0237&from=EN
17
Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022
Pillar 2: Advanced scientific training. Each DC will be assigned a combination of at least three elective, advanced
courses (made accessible to a wider audience), tailored to fellow’s experience, IRP and career plans (Table 1.3.c).
Many new modules will be created at the intersection of Finance, data science, AI, ML, explainability, blockchain
and sustainability which are not available in existing Finance PhD programs. Moreover, doctoral courses and summer
schools from our cooperation partners, the COST Action CA19130 , EIT Digital and ECMI will be opened up to all
recruited DCs.40
Pillar 3: Transferable skills training. DCs will undergo a tailored transferable skills development program, resorting
to courses that will be organised mostly via joint industry-academia courses (Table 1.3.d). Transferable skill training
is aligned with the 2018 Eurodoc report for early-stage researchers.41
Pillar 4: Training through secondments. Each DC will spend 18 months in industry and up to 6 months at one of
the world-leading research centres (European Central Bank, ARC, Fraunhofer Institute). DCs will significantly
expand their experience and skills by learning how Digital Finance is implemented both through traditional financial
intermediaries across Europe (DBA, RAI, SWE) and innovative Fintechs (ROY, CAR), how Europe is approaching
banking supervision, statistics, macroprudential policies and financial stability as well as international cooperation
by one of the leading central banks globally (ECB), how world’s leading applied research organisations prioritise key
future-relevant technologies and commercialise their findings in business and industry through innovative
developments and research excellence (ACR, FRA).
International conferences and other training programs. Network training events will often be combined with
larger events that include industrial partners, civil society, policymakers, and scholars from outside the network (e.g.
COST Action CA19130 with 240 researchers from 39 countries, ECMI with 170 research organisations across
Europe). This will help doctoral candidates build their research community and share their work beyond DIGITAL
(e.g. NeurIPS, AFA Annual Conference, EFA Conference, ECMI Conference, International R Conference,
Apply(conf), Data Reliability Engineering Conference, DeFi Conference).
Lab Training. DIGITAL has direct access to many data, analytics, blockchain, regulation & supervision labs. DCs
will receive training in utilising data, constructing ML and RL models, and validating results. Those are: University
of Twente and BFH’s Digital Finance labs, the ECB’s cloud-based digital virtual lab for fast-track innovation,
Fraunhofer’s virtual reality lab, optimization lab, computer graphics and visualisation lab, analytics lab, WULABS
for data analytics, ASE’s AI lab, Swedbank’s financial laboratory, and several more.
External scientific lecturers. DIGITAL features prominent external scientific lecturers from world-class institutes
who provide insights into the state-of-the-art in digital finance subdomains and excellent networking opportunities.
Those are Prof. A. Hirsa (Columbia University), Prof. C. Harvey (Duke University), Prof. O. Linton
(Cambridge University), Prof. J. Fan (Princeton University), Prof. M. Dai (Hong Kong Polytechnic
University), Prof. G. Stahl (Peking University HSBC).
Online courses. All course materials will be shared digitally and courses may be followed online (geant.org) if
appropriate . We will create new online content for long-term sustainability and re-usability.
Table 1.3.a Main Local Training Courses
# Course WP Description ECTS Month
1 Foundation of data science (BBU) 1 Introduction to a range of topics and concepts related to the data science 4 M12
process. It will cover the technical pipeline from data collection, to
processing, analysis and visualisation.
2 Introduction to AI for financial 2 Getting started with ML; Introduction to supervised and unsupervised 4 M12
applications (WWU) learning; deep learning and reinforcement learning. Explore how to use these
methods for financial applications (financial forecasting, credit risk, etc.)
3 The need for eXplainable AI: 3 Introduction to XAI methods; state-of-art models (LIME, SHAP, DiCE, LRP, 4 M12
methods and applications in counterfactual explanations, etc.). Challenges of classical methods.
finance (BFH) Introduction to methods suited for financial applications.
4 Introduction to Blockchain 4 Introduction to the blockchain technology, concepts such as mining, hashing, 4 M18
applications in finance (ASE) proof-of-work, public key cryptography, and the double-spend problem.
Overview of the design principles and challenges.
5 Sustainable finance (UNA) 5 Introduction to sustainable finance strategies. Overview of how these 4 M18
strategies can minimise organisational risk, create long-term business value,
and improve social and environmental impact.
40 https://www.cost.eu/actions/CA19130/, https://ecmiindmath.org/
41 http://eurodoc.net/skills-report-2018.pdf
18
Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022
6 Ethics applicable to digital 9 Introduction to Ethical Artificial Intelligence, with a specific focus on digital 4 M36
aspects (UTW) aspects.
Table 1.3.b Main Network-Wide Training Events, Conferences and Contribution of Partners
# Main Training Events & Conferences ECTS Lead Action
Month
1 Kickoff Meeting and Technical Training – University of Twente (NL) 3 UTW M12
[fellows, academic- and industrial supervisors, and representatives from associated partners]
Visits to the Finance Lab of the university will be arranged, where researchers will be offered several
training sessions like “Introduction to Digital Finance”, “Blockchains in Digital Finance”, and “Ethical
AI in Finance” by the UT faculty and additional consortium members.
2 Orientation Training Digital Finance – WU Vienna (AT) 3 WWU M15
[doctoral training combined with COST Action meeting]
This meeting will be organised in conjunction with a COST Action meeting, enabling fellows to meet
European researchers in Digital Finance outside the doctoral network. The meeting will be preceded by
a general five-day training course in Digital Finance.
3 Industrial Doctoral School on FinTech – EIT Digital (ES) 4 EIT M20
[open summer school hosted by EIT Digital]
EIT Digital will organise an open summer school in Madrid, tailored to doctoral candidates in FinTech.
Fellows will be able to interact with doctoral candidates in digital finance from outside the network. The
theme of the summer school is “Disrupting Finance with Digital Technologies”.
4 Regulation in Digital Finance Workshop – European Central Bank (EU) 2 ECB M24
[doctoral training combined with ECB site visit]
The workshop includes site visits to the ECB, providing direct immersion into the regulatory aspects of
digital finance, including topics such as AI bias, data sovereignty and digital currencies. Fellows will
attend a four-day training program. Two days are allotted to the topic of ‘Regulatory aspects of digital
finance’, conducted by regulatory experts from ECB.
5 Mid-term Review Event – Babeș-Bolyai University (RO) 1 BBU M30
[fellows, academic- and industrial supervisors, and representatives from associated partners]
The mid-term review event is a key event for the DIGITAL network that will bring together researchers,
industry, regulators and supervisors. The training component of the meeting is a short course on
“Sustainability in Digital Finance”.
6 Digital Finance Industry Event – University of Naples (IT) 2 UNA M36
[hosting industrial partners from within and outside the doctoral network]
Fellows will intensively interact with industry and orient themselves on potential digital finance careers
outside academia. Industrial partners from both within the network (CAR) and outside will provide
training on: “Fraud detection in digital accounting”, “Responsible AI in finance” and “Sustainable
digital finance”.
7 Training & Development Workshop – Kaunas University of Technology (LT) 2 KUT M40
[fellows, academic- and industrial supervisors, and representatives from associated partners]
In addition to an open conference on the topic of ML for Option Pricing, the event will also include
content training on “Designing digital finance tools” and a transferable skills training on
“Entrepreneurship in digital finance”.
8 Closing Conference – University of Twente (NL) 2 UTW M48
[fellows, academic- and industrial supervisors, and representatives from associated partners]
A selection of renowned keynote speakers from both academia and industry will speak at plenary
sessions. Fellows will also have the chance to speak with principal scientists and industrial partners,
reflecting on their work. The best project of the doctoral network, selected by the advisory board, will
receive a Best Doctoral Research award.
Table 1.3.c Advanced scientific courses
# Courses (E existing, WP Tutorial content Lead ECTS Month
N – new module)
1 Synthetic Data 1 Use of deep learning techniques (e.g., Generative Adversarial Networks) to ARC 4 M12
Generation for Finance generate synthetic financial data indistinguishable from real data. Use cases for
(N) synthetic data in AI training, e.g., fraud detection, crisis simulation.
2 Anomaly Detection in 1 Principles to detect anomalies. Discuss ways of handling data errors (e.g., human BBU 4 M18
Big Data (E, N) inspection, removing outliers, deploying AI to fill in gaps in data). Mapping of data
quality.
19
Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022
3 Natural Language 1 Combine computational linguistics and role-based modelling of human language ARC 4 M24
Processing with with statistical machine learning and deep learning models. Understand to use
Transformers (E, N) the most advanced transformers to perform advanced tasks.
4 Dependence Structures 1 Automatic detection of dependencies between arbitrary numbers of vectors. ASE 3 M30
in High Frequency Techniques for identifying patterns such as time-dependent trends, volatility
Financial Data (N) clustering, seasonality, and fat tails. Application of copulas and spectral measures.
5 Reinforcement Learning 2 Selection of learning algorithms relevant to digital finance applications. Deploying UTW 4 M12
in Digital Finance (N) RL for decision-making in areas such as trading, risk management, and fraud
detection.
6 Machine Learning in 2 Principles of machine learning in industry. Business assessment of automation CAR 4 M18
Industry (N) decisions. Practical implications of machine learning. Availability and costs of high-
quality data.
7 Deep Learning for 2 Build and train deep neural networks, identify key architecture parameters, BBU 3 M24
Finance (E, N) implement vectorized neural networks and deep learning. Analyse variance for DL
applications.
8 Data-Centric AI (N) 2 Empower SMEs in digital finance to deploy AI with limited datasets. Construct WWU 3 M30
high-quality samples to maximise training impact. Identify weak spots in data
quality.
9 Cybersecurity in Digital 3 Cybersecurity from perspectives of social behaviour, software and hardware. UTW 3 M12
Finance (N) Security of cloud services and compliance with EU regulations. Detecting and
preventing fraud.
10 A I Design in Digital 3 Overview of contemporary AI techniques in digital finance. Designing impactful AI ASE 4 M18
Finance (N) with explicit consideration for energy consumption, bias, explainability, and
fairness.
11 Barriers in Digital 3 Hurdles for society-wide adoption of digital finance. Design principles to WWU 3 M24
Finance Adoption (N) include genders, minorities and vulnerable groups. Fast-paced industry,
start-up climate, competition.
12 E xplainable AI in 3 Classification of white box- and black box models. Applicability of classical XAI BFH 4 M30
Finance (E, N) techniques in finance. Audience-dependent explanations. Emerging XAI techniques.
13 D igital Finance 4 Overview of the regulatory field in digital finance. Outlook to pending changes in ECB 3 M12
Regulation (E) EU regulations. Directions and focus points. Best practices for compliance and
monitoring.
14 H istory and Prospects of 4 Past developments in digital finance (including digital assets, algorithmic trading, UNA 3 M18
Digital Finance (N) AI) and trends for the next decade. Reflection on decentralisation. Reflection on AI.
15 B lockchains in Digital 4 Technical, financial and legislative principles of blockchain technology and its ASE 4 M24
Finance (E, N) (potential) applications in digital finance. Impact of decentralised finance.
16 D igital EIT Summer 5 Disrupting Finance with Digital Technologies. Reflection on the impact of FinTech EIT 4 M18
School (E, N) on society. Overview of latest advances. Case studies. Learning to write a business
plan.
17 G reen Digital Finance 5 Instill awareness of energy consumption and ecological footprint of digital KUT 3 M24
(E, N) finance. Techniques for energy-efficient algorithm training and deployment of
digital finance services. Trade-offs between performance and environmental
impact.
18 M ulti-Criteria Decision 5 Principles of multi-criteria decision making. Various techniques and concepts (e.g., FRA 3 M30
Making in Sustainable fuzzy set theory, analytical hierarchy process, preference modelling) to incorporate
Finance (E, N) multiple objectives, in line with ESG principles.
Our consortium will offer the following transferable skills courses:
Overview of commitment. Each DC will spend 18 months in the non-academic sector and up to six months in
the world-leading research centres and government agencies (ECB, BIS, ARC, Fraunhofer). Additionally, CAR
will be the beneficiary for two DCs. All non-academic partners have committed three to six employees to this
network, will make their in-house training facilities available to all DCs, and will contribute to research training
(30%), transferable skills training (60%) and our network-wide training activities in particular. In proportion to the
number of partners, the non-academic sector (including research centres and the European Central Bank) will
assume 50% of the responsibilities for leadership positions. Each non-academic partner has also agreed to provide
substantial additional hosting, training (Tables 1.3.a to 1.3.d), and infrastructure and laboratory resources
(Table 6). This is a substantial commitment from our non-academic sector that exceeds the norm for such set-ups.
Role in the training programme. Our non-academic partners play a crucial role in the training program (see Tables
1.3.c and 1.3.d) by exposing the fellows to the real financial industry, regulatory requirements, and technological
assets, enabling them to develop the scientific and industry-relevant skills needed to complete their PhD projects and
evaluate the research results from an industry perspective. The European Central Bank (ECB), the Bank for
International Settlements, financial institutions (Raiffeisenbank, DeutscheBank, Swedbank), and technology
companies (Cardo AI, Royalton Partner) provide fellows with a broad perspective to help transform digital finance.
All partners have large European R&D teams with academic ties. EIT Digital and its academic partners often train
doctoral candidates with industry, strengthening industry-academia ties.
Role in the research training. Non-academic partners shape the IRPs on which DCs work, providing data and
expertise in financial products and services, big data technologies, predictive modelling, and regulation. All partners
(co-)supervise doctoral candidates, contribute to all work packages, and have extensive experience (Table 6). The
European Central Bank will validate research and suggest strategic directions.
Experience of academic partners. The academic supervisors have (co-)supervised over 200 PhDs and many
postdocs and are frequently involved in PhD committees of prestigious international universities and as
organisers/ trainers at European doctoral schools. Each supervisor is a major partner in their doctoral programs,
published extensively in top journals with high impact, and is often (Co-)PI of large national and international
research projects with many PhDs. They (co-) lead prestigious international research consortia (Table 6). All
supervisors have extensive experience collaborating with industry, many with long work-experience, as e.g.
evidenced by FRA, an international research centre with annual research revenues exceeding 2 billion euros.
Non-academic partners. Our non-academic partners' supervisors are senior researchers or seasoned business
professionals. Most supervisors have PhDs and some hold professorships, and they co-supervise Master and Phd
students. All partners have large European R&D departments and are some of the most research-intensive
corporations across Europe, such as the European Central Bank.
Substantial academic credentials. Individually, the academic supervisors possess the profiles and credentials
necessary to guide and mentor doctoral candidates in their research projects and to support their training and
professional development. Given the synergy between the researchers and their ability to host doctoral candidates
during visits, secondments, and training schools, the network as a whole possesses even stronger supervision
capabilities and the collective quality to train doctoral candidates in an industrial doctoral network.
The proposed training methodology of DIGITAL will follow the principles for European Innovative Doctoral
Training42 and focuses on individual needs and aspirations of every recruited fellow, placing the Personal Career
Development Plans (PCDPs) at the centre of the training methodology. DIGITAL will comply with the European
Charter and Code43 and the European Code of Conduct for Research Integrity44, aiming to foster a constructive
environment to doctoral candidates conducting research and training45 in accordance with their IRPs.
The success of the supervision will be driven by this one-of-a-kind, academia-industry-leading network. We have
covered all pertinent financial areas, and all of our partners have substantial complementary skills and experiences,
42 Principles for innovative doctoral training:
https://euraxess.ec.europa.eu/sites/default/files/policy_library/principles_for_innovative_doctoral_training.pdf
43 The European Charter & Code for Researchers: https://euraxess.ec.europa.eu/jobs/charter
44 The European Code of Conduct for Research Integrity, 2017: https://allea.org/code-of-conduct/
45 Marie Sklodowska-Curie Actions Guidelines on Supervision, EC, 2021 (pdf).
21
Project: 101119635 — DIGITAL — HORIZON-MSCA-DN-2022
ranging from advanced business knowledge to theoretical academic research. Many supervisors have substantial
experience running industry-applied doctoral schools (Table 6).
DIGITAL will have a unified doctoral program across Europe, with three supervisors for each DC: an academic
supervisor, a non-academic supervisor, and one supervisor whose responsibility is to consider complementary skills.
The supervisors assist doctoral candidates with the creation, review, and implementation of their PCDP, offer
guidance on their IRP, and provide professional and scientific guidance.
All doctoral candidates' supervisors will evaluate their work, provide feedback on their progress and research, be
available for scheduled meetings, and actively participate in their training. The surveys conducted by DCs will help
assess and enhance the program. Supervisors are routinely accessible and dedicate significant time to doctoral
education. The academic supervisor is responsible for overseeing the IRP's scientific contribution, career advice
for doctoral candidates, and the activities' schedule.
The non-academic supervisor introduces them to relevant departments, facilitates their practical problemsolving
skills, broadens their network, and assists them in applying their knowledge to real-world issues. The third
supervisor is responsible for ensuring that DCs acquire soft skills and apply their knowledge in a variety of
application areas. All supervisors have complementary expertise (Table 6). Our complementary partners have worked
on many research projects together for many years, for the benefit of our DCs. To ensure representation, each
supervisor has one vote in the SB. We also have industry-experienced supervisors from large international
research centres (e.g. FRA with more than 2bn EUR research revenues per year). All supervisors of DCs are
indicated in Table 6. 18-months industry secondments in Table 3.1.c correspond to the non-academic supervisor.