The conference brings together Machine Learning experts from the Financial Services space to lead key discussions surrounding some of the biggest topics surround the industry at the moment. Bringing together senior executives from private and public banks, payment services and insurance companies to explore the emerging themes when deploying Machine Learning models within their day to day services.
4. COST REDUCTION
Lower the compliance
burden with better
productivity due to process
automation
FAST INSIGHT
Provide real time
insight for decision-
making reducing the
time to market
BETTER
COMPLIANCE
Produce error-free
outcomes while
scaling up the scope
IMPROVED SERVICES
Optimize customer
journeys and anti-fraud
security
5. Machine learning will
improve performance when
business rules are fully
understood and the
governance is supported by
investment in
infrastructure
7. Promises to tangible
benefits?
Invest in pilots of suitable
business cases
Historical data to future
trends?
Monitor changes in
conditions and reduce noise
Discrimination to learning?
Prevent algorithm to
deteriorate over time
Restrictive to flexible
models?
Balance supervised and
unsupervised learning
8. Risks of machine learning in financial services
• Biased assumptions may be amplified to discriminate
customers
• Inaccurate algorithms may provide wrong advise to
customers and account officers
• Profiling may breach privacy compliance requirements
• Errors may occur due to the lack of skills and knowledge
• New threats and attacks may manipulate algorithms
• Algorithms may be inaccurate for unanticipated situations
and extreme cases
• GDPR requirements may be breached by cross-border data
transfers
9. Tips
• Request validations and revalidations of the models
• Test the models with segmented diverse data sets
• Increase the testing when unstructured data is used
• Decrease the risk tolerance for automatic trading instructions
• Assess risks in new projects including those for customers
• Start from non-core functions and non-sensitive cases
• Use parallel implementations
In general, caution
• Treat machine learning as an emerging technology to be
rolled out depending on each case
10. Are we ready?
• In 2019, nobody thought that financial institutions could be
ready for fully remote operations
• There are more needs to understand fraud and compliance
risks at process and country level
• The maturity of machine learning depends on each process
• Low complexity and high volume processes are best
suitable for machine learning
• Increasing regulations and orders prevent processes to be
stable
11. • Get consents from clients
• Address the risks for
discrimination
AI ethics
• Supervise outcomes
• Supervise AI systems with
automatic learning capabilities
• Address the limitations
14. Machine learning
governance
Reliable outcomes of models
• Apply ethical principles
• Secure environment for
development
• Validate data sources
• Monitor changes in the
environment impacting the
underlying statistical properties
16. CHECKS
Automatically retrieve
documents and internal and
external data for
investigations triggered by
alarms
DOCUMENTATION
Automatically upload
validated
documentation into
compliance systems
STANDARIZATION
Check integrity to
avoid
inconsistencies,
and missed
documents
SCREENING
Automate matches against
sanctioned and listed
entities
17. TEMPLATES
Automatically address the
same type of repeated alerts
with a template of response
ON/OFF BOARDING
Automatically open
and close customer
accounts in different
systems with
attached
documentation
REPORTING
Automatically retrieve,
standardize and
aggregate data from
structured and non-
structured sources
KYC and CDD
Automate from the
document scanning to data
interfacing
18. Risks of robotics in anti-money laundering
• Maintenance over costs and non-compliances may occur
due to frequent changes, inconsistencies and errors in the
operational processes
• Errors in executing bots may be due to low quality of data
sources, in particular, legacy systems
• Incompatibilities to adjust to new requirements may be
caused by obsolesce of the robotics solution
• Fully automated processes may not be possible due to
complexity and changes in the operational activities and
requirements
19. Recommended controls for robotics
• Segregate access profiles for creating, executing, and
maintaining bots in development, testing, and production
environments
• Centralize and limit the rights to configure the robotic solution
• Monitor and respond to alarms on load balances of the server
• Increase approvals and testing for change management,
including using clients
• Review logs on critical activities for both users and bots
• Restrict physical access to the control room with the robotic
server in on-premise settings
• Include the robotic server in the penetration tests
15 minutes for presentation
ooking to the Emerging Markets for the Future of Machine Learning Innovation
Due to the impact of Covid the emerging markets in China, Singapore and India have taken major strides in their commitment to advancing their ML processes
Exploring the advances in the payment systems in countries such as India utilizing Machine Learning functions to streamline efficiency
Looking towards the alternative lending platforms available in emerging markets to allow access to underserved communities
Assessing the impact on workers in industries whose roles may be automated thus exacerbating socio-economic issues within vulnerable populations
And the debate is as follows:
Debate: Is Machine Learning Mature Enough to Successfully Implement in Financial Institutions
While everyone is quick to jump onto the Machine Learning trend, is it really safe to implement within the financial services sector with so many issues surrounding the regulatory and ethical side of utilizing machines to make human decisions?
Overcoming the issues faced when explaining outcomes that may be discriminatory which can damage a company’s reputation
Is Machine Learning really needed to automate financial processes or does the negativity around ethical considerations enough to reconsider?
Can regulatory bodies ever be confident enough in the decisions made by the machines to allow ML to really progress in financial services?
Looking towards ensuring transparency in the models decision making process to determine if it is suitable for deployment in financial decisions
Looking to the Emerging Markets for the Future of Machine Learning Innovation
Due to the impact of Covid the emerging markets in China, Singapore and India have taken major strides in their commitment to advancing their ML processes
Exploring the advances in the payment systems in countries such as India utilizing Machine Learning functions to streamline efficiency
Looking towards the alternative lending platforms available in emerging markets to allow access to underserved communities
Assessing the impact on workers in industries whose roles may be automated thus exacerbating socio-economic issues within vulnerable populations
And the debate is as follows:
Debate: Is Machine Learning Mature Enough to Successfully Implement in Financial Institutions
While everyone is quick to jump onto the Machine Learning trend, is it really safe to implement within the financial services sector with so many issues surrounding the regulatory and ethical side of utilizing machines to make human decisions?
Overcoming the issues faced when explaining outcomes that may be discriminatory which can damage a company’s reputation
Is Machine Learning really needed to automate financial processes or does the negativity around ethical considerations enough to reconsider?
Can regulatory bodies ever be confident enough in the decisions made by the machines to allow ML to really progress in financial services?
Looking towards ensuring transparency in the models decision making process to determine if it is suitable for deployment in financial decisions