Deep Learning in Finance for Pension Funds with Examples

Einleitung/Dauer

Deep Learning supports Solvency Requirements for Pension Funds and ensures the Guarantee of Benefit Payments at any time. It is particularly well-suited for time series forecasting related to inflation, yield curves and asset allocation returns in order to prepare and review these. We use the Long short-term memory (LSTM) method, a type of recurrent neural network (RNN), as well as other neural networks.

Deep learning is a type of machine learning that uses multi-layered neural networks. The libraries programmed with Python are fascinating areas of research because they help to verify time series forecasts and to understand how long it might take for the pension fund to reach the target value of the investment fluctuation reserve based on the current situation.

RNN-based models, particularly LSTMs, are increasingly being used to capture complex spatio–temporal dependencies, while hybrid architectures combine convolutional and recurrent components (i.e., CNN-LSTM). Researchers have developed hybrid models that further improve prediction accuracy, which is very important for financial forecasting.

Preliminary Programme
Monday, 29 June 2026
10:00-11:00 Topics:

• Principles of Deep Learning in the financial forecasting.
• Explanations of recurrent neural network (RNN) model LSTMs
• Useful Python libraries for Deep Learning
• Examples forecasting with Pictet indices (daily and monthly return data) and UBS Pension Funds Indices (monthly return data)

11:00-11:15 Break
11:15-12:15 Topics:

• Examples forecasting indices for equities (download from the financed library)
• Explanation how LSTM models help to produce LLMs for pension funds
• Explanations of how to verify forecasts based on different approaches prepared with historical data
• Summary

All the above times are given in CEST (Central European Summer Time).

Vorgehensweise und Ziele

The annual financial statement of a pension fund shows all important parameters of the liabilities as well as all types of reserves. Deep Learning in Finance helps to forecast for portfolio returns, inflation as well as yield curves and to understand how the investment fluctuation reserve might develop. Based on this analysis, the annual financial statement presentation could be prepared for the members of the Board of Trustees, helping them to make final decisions on the expected benefit payments and to understand how this kind of analysis could be done.

Teilnehmer

The web session is suited for pension fund actuaries and actuarial professionals, IT-developers of pension fund software tools who are directly or indirectly involved in actuarial and investment consulting for pension funds and collective foundations with occupational provisions. Additionally, these topics could be useful as well to members of the pension fund board of trustees, pension fund managers, and pension fund auditors.

Technical Requirements
Please check with your IT department if your firewall and computer settings support web session participation (the programme Zoom will be used for this online training). Please also make sure to join the web session with a stable internet connection.

Dozierende

Dr Ljudmila Bertschi
Ljudmila is a qualified member of the Swiss actuarial association (SAV/SAA) and an accredited pension actuary of the Swiss chamber of pension fund experts (SKPE). She holds a PhD in phys.-math. from the MSU and has been working in pension fund consulting for various Swiss and international consulting firms and insurance companies for about 20 years. She conducted a research study for the Federal Office of Social Security (2015), has authored many publications and presentations for international conferences and has given training lectures for the Swiss chamber of pension fund experts (including on liability forecasting with Markov chains). Two EAA-Workshop were prepared and presented by Ljudmila together with Mauro.

Dr Mauro Triulzi
Mauro is a qualified member of the Swiss actuarial association (SAV) and holds a PhD in mathematics from ETH Zurich. He has been working as a developer of actuarial tools for around 20 years and implemented nested stochastic modelling for pension fund liabilities including mortality rates for ALM studies. He is currently developing various actuarial tools for local and international accounting valuations as well as pension fund administration services. He has prepared presentations for international conferences as well as two EAA workshops together with Ljudmila.

Sprache/Kurztitel

The language of the web session will be English.

CPD Credits
For this web session, the following CPD credits are available under the CPD scheme of the relevant national actuarial association:

• Austria: 2 points
• Belgium:  2 points
• Bulgaria:  3 points
• Croatia: individual accreditation
• Czechia:  2 hours
• Denmark 2 credits
• Estonia:  2 hours
• Finland: 2 points
• France:  12 points
• Germany:  2 hours
• Greece: 3 points
• Hungary:  2 hours
• Iceland: 2 credits
• Ireland: 2 hours
• Italy: individual accreditation
• Latvia: 2 hours
• Lithuania: 2 hours
• Netherlands:  approx. 2 points (individual accreditation)
• Norway: 2 points
• Poland: 2 hours
• Portugal: 2 hours
• Serbia: 2 hours
• Slovakia:  individual accreditation
• Slovenia:  individual accreditation
• Spain: CAC: 2 hours, IAE: 2 hours
• Switzerland:  individual accreditation
• USA: SOA (Section B): up to 2.4 hours

No responsibility is taken for the accuracy of this information.