INFINITECH - GA 856632 Project Overview

1This project has received funding from the European Union’s horizon 2020 research and innovation programme under grant agreement no 856632

Flagship initiative for Big Data in Finance and Insurance

Ernesto Troiano, GFT Italia

INFINITECH - GA 856632Project Overview



Tailored IoT & BigData Sandboxes and Testbeds for Smart, Autonomous and Personalized Services in the European Finance and Insurance Services

Ecosystem

• Response to the CALL H2020 ICT-11-2018 Innovation Action

• Grant Agreement INFINITECH # 856632

• Project Budget: 21.080.482,00 €

• EC Grant: 15.870.480,00 €

• Started 1st October 2019 Ending 31st December 2022

• Duration 39 months

• 48 (45+3) Participants (Industry, Banks and Insurance, Academia, FinTechs)

• Reporting 1st Period until 31st March 2021



Historically, the financial and insurance services sectors are “conservative” (resistant to digital transformation and technology disruptions)

Digital Technologies (AI, BigData, IoT) and recent regulatory developments (PSD2) are accelerating the transformation of the financial and insurance services industry

The majority of digital transformation applications for the finance and insurance sectors are data-intensive and leverage technologies on fragmented data sources

Endless quest for automation, cost-effective applications and disruptive business models exploiting Big Data and AI in the frameworks of Directives and Regulations

Background



• AI architectures in their infancy

• No-adequate support for real-time AI use cases• e.g., detection of fraudulent transactions on the fly

• Limited Data Sets for Training Algorithms

• Data Bariers (“Silos”)

• Lack of testbeds & experimentation resources (Data Assets & Sandboxes”)

• Business Models not validated

• Complex Regulatory Environment

INFINITECH Drivers & Motivation



Project Goals Demonstrate innovation potential of AI, BigData, IoT, and Blockchain in Digital Finance• Novel concepts and

approaches• New products, services or

business and organisationalmodels

• Enhance competitiveness of Financial Organizations and FinTechs

• Create new market opportunities based on novel business models

• Accelerate growth of companies (SMEs, FinTechs)

• Capacity and Capabilities in execution - Complementarity of the participants

• Validation in Pilots Covering a very broad range of the sector (e.g., from Personalized Asset Management to AML and Fraud Detection

INFINITECH is a joint effort of leaders in ICT and finance towards:• Lower the barriers for BigData/AI driven innovation

• Boost regulatory compliance - Stimulate additional private investments• Provide a Broad Coverage of Use Cases in the Finance and Insurance Sectors

Innovation Market Relevance Execution



INFINITECH MissionHelp Financial Organizations and FinTech enterprises to Democratise Digital

Finance through highly personalized and trusted services leveraging cutting edge BigData, AI and Blockchain technologies.

Democratization

• Lower Adoption Barriers for Customers (Trust, Access, Regulatory Compliance)

• Facilitate Financial Organizations to use and fully leverage BigData, AI, Blockchain (e.g., novel technologies, organizational Models, business models etc.)

Personalization

• Personalized Finance Management

• Personalized Investments & Asset Management

• Customer Centred Analytics

• Personalized Usage Based Insurance (Healthcare, Vehicles)

• Fraud Detection

• Anti-Money Laundering

Cutting Edge Technology

• Large Scale Data Integration & Interoperability

• Real-Time & Incremental Analytics

• Novel AI/ML Models and Algorithms for Digital Finance

• Blockchain Asset Tokenization and Graph Analytics

• Etc.



INFINITECH Overview

Flexible Reference Architecture to map the 15 Pilots

Innovation



INFINITECH Approach Innovation



The INFINITECH Team



INFINITECH WorkplanWP1 Project Management and Administration

WP9 Dissemination, Exploitation and Standardization

WP2

Vis

ion

and

Spec

ifica

tion

s fo

r A

uton

omou

s, In

telli

gent

and

Pe

rson

aliz

ed S

ervi

ces

in

Fina

nce/

Insu

ranc

e

WP5 Open Autonomous Quality Services

Engineering and Processes

WP3 BigData/IoT Data Management and

Governance for SHARP Services WP6 Tailored

Sandboxes and Testbeds

for Experimentati

on and Validation

WP7 Large-Scale Pilots of

SHARP Financial and

Insurance Services

WP8 Pan-European

Multi-Sided Market

Platform and VDIH

WP4 Interoperable Data Exchange and Semantic

Interoperability

Execution



The INFINITECH Pilots (Finance)

Personalized Retail and Investment Banking Services

• KYC & Customer-centric Data Analytics

• Personalized Portfolio Management & “Democratization” of Investment Banking

• Smart & Personalized Pocket Assistant for PFM

• BFM tools delivering a Smart Business Advise

• Personalized Closed-Loop Investment Portfolio Management for Retail Customers

Smart, Reliable and Accurate Risk Assessment

• Invoices Processing Platform for a more Sustainable Banking Industry

• Real-time Risk Assessment in Investment Banking

Predictive Financial Crime and Fraud Detection

• Operation Whitetail – Avoiding Financial Crime

• Platform for Anti Money Laundering Supervision

• Analyzing Blockchain Transaction Graphs for Fraudulent Activities

• Real-time cybersecurity analytics on Financial Transactions’ BigData

Market



Personalized Usage-Based Insurance Pilots

• Personalized insurance products based on IoT connected vehicles

• Real World Data for Novel Health-Insurance products

Configurable and Personalized Insurance Products for SMEs and Agro-Insurance

• Alternative/automated insurance risk selection -product recommendation for SME

• Big Data and IoT for the Agricultural Insurance Industry

The INFINITECH Pilots (Insurance) Market



Project Status



The INFINITECH Achievements

15 Active Pilots

15 Testbed Infrastructures (10 Incumbent + 5Infinitech Cloud)

15+ Prototypes & Demonstrators

40+ Technological Tools/ Assets

60+ Building Blocks/Components

1 Reference Architecture and 15Blueprints (1 per pilot)

84 Deliverables Submitted (more than 4000 pages)

1 Marketplace + 1 Hub

350 Stakeholders Engaged + 1 Task Force on Digital Finance

11+ Workshops (and innumerable project presentations)

10 GA/Meetings (three face to face before pandemic)

100+ Telcos



• Provide a Reference Architecture to be used as a blueprint for integrating and deploying the components that comprise an IoT/BigData/AI application

• Unified data management layer enabling access over fragmented silos, providing operational and analytical processing capabilities on real data

• Framework for data governance and tools for regulatory compliance

• Semantic Interoperability framework for data exchange among finance/insurance institutions

• Library of ML/DL algorithms

• Definition of tailored sandboxes for automatic deployment of integrated solutions

• Validation of INFINITECH technologies by pilots

• Definition of specific business plans and business methodology

Key Objectives



The INFINITECH WAY



• A set of Technologies (BigData, AI, Blockchain) for the industry

• A Reference Architecture

• A set of components at high level of maturity (TRL)

• Infinistore & Query Language/DataGovernance Tools/Semantic Engine/NFT/Machine Learning

• A Development and Deployment Methodology

• A Reference Infrastructure

• A set of Blueprints for Development, Implementation

• A set of demonstrators compliant with INFINITECH

• A set of channels for Communication/ Dissemination/ Exploitation (Marketplace / VDIH)

• A Reference for Business Models and Exploitation Plans

THE INFINITECH WAY Innovation



INFINITECH Reference Architecture

R.A. & Building BlocksInnovation

BDVA/DAIROCompliant


Flagship initiative for Big Data in Finance and Insurance A Workflow ExampleExecution



“Unified data management layer enabling access over fragmented silos,providing operational and analytical processing capabilities on real data”

• INFINISTORE as the data management layer• HTAP provision

• High rate data ingestion

• Polyglot extensions for integrated access of data “silos”

• Integrated query processing of streaming and static data

• One patent to be filed for LXS in the next period

Unified Data Management



“Framework for data governance and tools forregulatory compliance”

• 2 Anonymization tools using new utility metrics

• Digital User Onboarding Service, using an improved face recognition process

• Mapping data regulations with the identified INFINITECH technologies and cross check with actual directives/laws

• Preliminary overview of the general regulatory compliance tool for INFINITECH, using a Data Protection Orchestrator (DPO) tool

Data Governance for Regulatory Compliance



“Semantic Interoperability framework for data exchangeamong finance/insurance institutions”• INFINITECH Semantic Interoperability Framework (FIBO, FIGI, LKIF

Compliance))

• INFINITECH Methodology for Ontology Engineering and Prototyping

• Implementation of connectors and transformers for ingesting pilot-specific data

• Definition of INFINITECH Graph Data Model

• Online Tools for Data annotation and Semantic Interoperability

• Semantic Stream Analyzer Middleware/Engine (SeSA-me) for stream processing

Semantic framework



“Provision for advanced analytical capabilities”

• Declarative framework for real-time analytics• Introducing ONLINE AGGREGATES• Patent filed

• Provision of incremental scans

“Library of ML/DL algorithms”• Provision of more than 10 ML/DL techniques for digital finance

problems.• Incorporation of Psychometric and BetaRecys ML/DL engines• Integration of 2 machine learning data analytic platforms• Framework for definition and deployment of ML-based

microservices

Advanced data analytics and AI algorithms



“Definition of tailored sandboxes for automatic deployment ofintegrated solutions”

• Reference testbed available for testing and open pilots

• INFINITECH way for deploying tailored sandboxes• Deployment recipes for INFINITECH building blocks

“Validation of INFINITECH technologies by pilots”

• Primary results from pilots already utilizing INFINITECH technologies

• Applications deployed using INFINITECH recipes and building blocks

• Integrated solutions deployed in reference testbed, using cross WP technology assets

Tailored sandboxes and pilot deployment



INFINITECH Testbeds & Sandboxes

INFINTECH technological

building blocks

Data assets

Regulatory Compliance Tools

ML/DL algorithms

Open API

Execution



Marketplace, VDIH and ISA

ISA will be one of the main vehicles for engaging relevant stakeholders• Stakeholder ecosystem• Structured communication• Privileged access to information, assets

and services• Collaboration with R&Is• Newsletter with focused information

Market



INFINITECH Impact Creation Market

I m p a c t

•New Services and Products for end users and customers

•Ecosystem and Community Building around Marketplace and VDIH

•Links to Prominent Communities (DAIRO, GAIA-X, EBA,…)

•Disruptive Approach for Industry (Financial Insurance)

Leverage the expertise and services of the consortium

Services differentiated bycategories

Agile search with filters

Based on Marketplace infrastructure

Specific layer inside thePresentation Layer

VDIH

Network and collaborations

Innovation Support

Hackathons

Courses



[email protected]

Thank you



Cate. Smart and Reliable Scoring, Risk and Service Assessment

Pilot #1 Invoices Processing Platform for a more Sustainable Banking Industry

Description Partners

Develop, integrate and deploy a data-intensivesystem to extract information from notary invoices toestablish a sustainability index of notary servicesbased on the number of physical copies issued, thatwill be used by the bank.

● BANKIA

○ Data custody provider and financial services expert

○ Access to datasets (notary invoices) and infrastructure

● GFT

○ System and solution integrator

○ Big data analytics, AI expert

○ Sustainability scoring definition

● FBK (Fondazione Bruno Kessler)

○ Data science expert

○ AI algorithms development and data processing

○ Pilot automation

● RB (Report Brain)

○ Text-analytics algorithm/ solution

○ High-data-velocity processes.

● INSO (Insomnia Digital Innovation Hub)

○ participates as business and development advisor



Categ Smart and Reliable Scoring, Risk and Service Assessment

Pilot #2 Real time risk assessment in Investment Banking


Provide detailed risk information in real time and

produce aggregated reports.

Implement risk assessment and monitoring procedurefor two standard risk metrics: • VaR (Value-at-Risk) • ES (Expected Shortfall)

Supported by real-time sentiment analysis of newsarticles from different sources

● JRC leader and pilot owner

● INNOV support proxy working on analytics

● GFT pilot design and prototyping



Categ Smart and Reliable Scoring, Risk and Service Assessment

Pilot #15 Open Inter-Banking Pilot


Pilot objectivesThe Inter-Banking Open pilot, is the result of an Open Call to shared

business pains among Banks, and its objective is to develop a

solution that could address and tackle such pains in a pre-

competitive environment. Several use cases have been gathered,

evaluated and discussed. The final result is an intelligent system

capable of reading, analyzing, filtering and organizing the Banks'

documents by developing and exploiting an innovative taxonomy.

The pilot aims to implement the prototype of a solution based

on Machine Learning and Natural Language Understanding

paradigms. This prototype will start from the analysis of a

subset of process operating documents to attempt the

classification of the information contained in them, used by

Italian banks to build their business glossary and to support the

Enterprise Architecture Modelling

● ABI Lab : Italian banks: ABI Lab is the Banking Research

and Innovation Centre founded and promoted by the

Italian Banking Association (ABI)

● (GFT+HPE) technical partnerts supporting the pilot




● This pilot would examine how banks and fintech(s) in collaboration with research organisations and NGOs can develop an AI driven capability using transactional data generated by the financial activities that identifies money-related profiles based on the transactional data generated. Data profiles e.g. from social media then can be associated to human profiles base on their financial activity. These profiles will be built into the available AI engine and will be combined with existing technology and data sourced from the TAH human trafficking platform. The results will produce a complete picture of people profile, people trafficking routes and the corresponding money flows back to the criminal organizations

Bank of Ireland

TAH

BPFI

NUIG

Categ Personalized Retail and Investment Banking Services

Pilot #3 Collaborative Customer-centric Data Analytics for Financial Services




Develop and adapt within Privé Managers Wealth

Management Platform an Optimization algorithm, as

well as improving and expanding its capabilities as an

artificial intelligence engine to aid investment

propositions for retail clients.

● PRIVE

○ Pilot Owner

● RB (ReportBrain)

○ Analysis of financial news data to quantify the current

economic market sentiment.


Pilot #4 Personalized Portfolio Management (“Why Private Banking cannot be for everyone?”)



Categ Personalized Retail & Investment Banking Services

Pilot #5b Business Financial Management (BFM) tools delivering a Smart Business Advise


Enable SMEs to better understand and

be in control of their business/financial health.

Doing so by providing to SMEs smart personalized

insights generated through a suite of Business

Financial Management (BFM) tools leveraging AI.

● BOC (Bank of Chyprus)

○ will lead this pilot and represents the end-user,

testbed and data provider for this pilot. BOC will set

the business requirements and will support the

development of the predictive AI-based BFM tools

● UPRC (University of Piraeus Research Center)○ AI – Machine Learning Models/Development

● GFT○ Integrator

● (CP) Crowdpolicy○ Chatbot




Pilot objectives• The goal is to create a system for personalized investment

recommendations for the retail customers of the bank

• NBG will leverage large customer datasets and large volumes of

customer-related alternative data sources (e.g., social media,

news feeds, on-line information)

• NBG’s business motivation behind developing this pilot is:

○ Enhanced Productivity

○ Better Advice for Investment

○ Increased Trading Volumes

● NBG (National Bank Greece)

○ as main data provider and product owner. NBG will

provide the sandbox and data for experimentation.

● CP

○ as integrator of INFINITECH technologies and

Machine Learning / Deep Learning algorithms.

● RB

○ as provider of AI technology over open and

alternative sources (social media, news).

● PRIVE

○ as business expert in investment banking and asset

management recommendations, including synergies

with its own pilot


Pilot #6 Personalized Closed-Loop Investment Portfolio Management for Retail Customers



Categ Financial Crime and Fraud Detection

Pilot #7 Advanced Financial Crime Risk Model and Scoring


Improve the detection of Financial Crime usingrefreshed customer data and AI based analysisof customers‘ transaction behavior

● CXB (CaixaBank)

● FTS (Fujitsu)

● GFT

○ system integrator

● FBK (Fondazione Bruno Kessler)

○ research partner

● INSO – advisor




Pilot #8 Platform for Anti-Money Laundering Supervision (PAMLS)


Develop a platform that will improve the effectiveness of theexisting supervisory activities in the area of AML/CTF by processinglarge quantity of data (Big Data) owned by the BOS and othercompetent authorities (e.g. FIU).

● BOS (BANKA SLOVENIJE)

● JSI (Jožef Stefan Institute)




Blockchain crypto currencies and tokenized assetsthat are obtained fraudulently can go through varioustransfers on the blockchain and end up as stable coins( e.g. USD, EUR, TRY tokens) in different jurisdictions.As a result, it is possible that a company that acceptsstable coins, is paid by stable coins that can be tracedto addresses involved in fraudulent activities.Holding stable coins that originated from fraudulentor sanctioned addresses can be risky for thecompany. Hence, construction of the massiveblockchain transaction graph and its analysis isnecessary to detect frauds .

● AKTIF (Aktif Bank)

○ Responsible for user interfaces and

regulations and banking services.

● BOUN (Bogazici University)

○ Responsible for HPC software

development for big blockchain data and

parallel graph analysis.


Pilot #9 Analysing Blockchain Transaction Graphs for Fraudulent Activities




The pilot will build a data analytics platform to help Nexi AML team to discover, monitor and analyze suspicious scenarios related to money laundering through digital card payments.Nexi, as the Italian paytech leader, owns and manage a large big data ecosystem, which includes information regarding cardholders, merchants, organizations, and digital payment authorizations and transactions.The Nexi AML team purpose is to preside anomalous scenarios linked to money laundering, adhering to European AML regulatory compliance policies, by notifying detected cases to the Italian Financial Intelligence Unit (FIU).

NEXI is the end-user, tested and data provider for this pilot. GFT will act as the integrator of INFINITECH technologies


Pilot #16 Data Analytics Platform to detect payments anomalies linked to money laundering events.



Cate

g

Personalized Usage-Based Insurance Pilots

Pilot #11 Personalized insurance products based on IoT connected vehicles


Improve the risk insurance profiles using the

information collected by connected vehicles and

applying IoT, HPC, Cloud Computing and Artificial

Intelligence technologies: drivers’ classification and

fraud detection.

● ATOS

○ Connected Car Platform

○ IA Platform

● CTAG

○ On board units

○ Real-Time/Historical data

○ 80 vehicles during 4 by day

● Gradient

○ Anonymization Service

● Dynamis

○ Requirements

○ insurance company’s data




● Improve the risk insurance profiles○ using the information collected by

activity trackers & questionnaires

○ and applying IoT & ML technologies

● SiLo

○ Leader

○ ML services

● iSPRINT

○ Data collection platform (Healthentia)

○ ML services

○ Pilot user monitoring

● RRD

○ User aspects: pilot setup and running

● Gradiant

○ Anonymization Service

● Dynamis

○ Requirements

○ Domain knowledge

Categ Personalized Usage-Based Insurance PilotsPilot #12 Real World Data for Novel Health Insurance




Pilot objectives

The Pilot#13 will implement an automation of the subscriptionprocess helps the insurance company reduce costs. In addition, beingable to verify that the data entered is correct with a doubleverification avoids possible errors in the cost of the insurancepremiumThe monitoring and identification of real-time risk changes allowsthe company to know if the insurance cost really corresponds to thereal risk of the SME or if it should increase or decrease it to adapt itto its current situation.

● Wenalyze

○ Big Data Analitics platform

● LX (LeanXcale)

○ Datastore provider

○ Poliglot provider

● RB (ReportBrain)

○ participates a solution expert with expertise in text-

analytics and credit rating.

Categ Configurable and Personalized Insurance Products

Pilot #13 Alternative/automated insurance risk selection - product recommendation for SME




Define, structure and pilot test specific services

for the Agricultural Insurance sector in order to

better protect agricultural assets by evaluating

risks in a data-driven way:

(1) Mapping of risks related to agriculture in

predefined markets

(2) The prediction and assessment of weather

and climate risk probability

(3) Damage assessment calculator for insurance

companies.

● AGRO

○ Service-Infrastructure

○ EO Data and Weather intelligenc services

● GEN

○ Process evaluation

○ Pilot implementation with insurance clients

○ Dissemination of pilot results

Categ Configurable and Personalized Insurance Products

Pilot #14 Big Data and IoT for the Agricultural Insurance Industry

Date post:	02-Nov-2021
Category:	Documents
Upload:	others
View:	3 times
Download:	0 times

INFINITECH - GA 856632 Project Overview

Documents