---

The conference took place at Bayes Business School, City St George’s, University of London, 19 - 20 June 2025

Programme & slides

• City, University of London, Northampton Square, London EC1V 0HB
• Download conference abstract and programme booklet
• Slides will be made available here over the coming days.

19 June 2025

08:30 - 09:00 Registration

09:00 - 09:15 Welcome (Room B200)

09:15 - 10:15 Keynote 1 (Room B200) (Chair: Andreas Tsanakas)

Magdalena Ramada (WTW): Emerging Technologies - Current Impact on the Insurance Value Chain

10:15 - 11:15 Regular Session 1

Stream 1 Room B200: Data (Chair: Markus Gesmann)

• Iftekhar Khaled, Chris Halliwell (Markel): NLP for data granularity improvement
• Mick Cooney (Describe Data): Analysing ship utilisation using AIS data
• Manuel Caccone (Italian Society of Actuaries (ISOA)): Application of NLP models in loss modeling for actuarial science

Stream 2 Room BG02: Fairness (Chair: Karol Gawlowski)

• Charlotte Jamotton (Université Catholique de Louvain): A multivariate energy distance approach to premium fairness adjustment
• Mathias Lindholm (Stockholm University): Sensitivity-based measures of discrimination in insurance pricing
• Marie-Pier Côté (Université Laval): A scalable toolbox for exposing indirect discrimination in insurance rates

11:15 - 11:40 Coffee break

11:40 - 12:40 Lightning Session 1

Stream 1 Room B200: Natural Language Processing (Chair: Ioannis Kyriakou)

• Amin Karbassi (AXA XL): Track trending topics in insurance for emerging risk identification
• Claudio Giorgio Giancaterino (Towards Innovation Lab): Harnessing conditional generative models for synthetic non-life insurance premium data
• Pratyush Singh (Swiss Re): Enhancing contract wordings analysis with generative AI: A timeline of efficiency and accuracy
• Matej Otcenas (Swiss Re): Advancing Claims Document Processing with LLMs: From PoC to Production
• Malgorzata Śmietanka (UCL): LLMs for claims processing: A fully local and compliant solution
• Daniel Jakobi (Markel): Identifying similar insurance claims using text-based vector search

Stream 2 Room BG02: Climate (Chair: Mathias Lindholm)

• Hirbod Assa (University of Essex): Drought parametric insurances by a two-step machine learning approach under climate change scenarios
• José Luis Vilar-Zanón (Complutense University of Madrid): The influence of climate change on insurance sustainability: Evidence from Spanish agricultural insurance
• Yubo Rasmussen (Heriot-Watt University): A sub neural network approach for forecasting climate-related claim costs in property insurance
• Despoina Makariou (University of St.Gallen): Mitigating systemic risk in catastrophe insurance: The role of human judgment in model diversification
• Yushan Liu (Institute Polytechnique de Paris): Meta-modelling paths of simple climate models using Neural Networks and Dirichlet polynomials: An application to DICE
• Francesco Ungolo (University of New South Wales): An augmented variable Dirichlet process mixture model for the analysis of dependent lifetimes

12:40 - 13:40 Lunch

13:40 - 14:40 Regular Session 2

Stream 1 Room B200: Pricing Methods (Chair: Bernard Wong)

• Robert Carruthers, Shane Murphy (Ki Insurance): Predicting the full D&O insurance tower with securities class action data
• Pierre-Olivier Goffard (Université de Strasbourg): Market-based insurance ratemaking: Application to pet insurance
• Salvatore Scognamiglio (University of Naples Parthenope): The credibility transformer

Stream 2 Room BG02: Mortality (Chair: Pietro Millossovich)

• Carlos Arocha (Arocha & Associates GmbH): Climate-enhanced pricing: Using gradient boosting machines to personalise life insurance rates
• Jens Robben (University of Amsterdam): Granular mortality modeling with temperature and epidemic shocks: A three- state regime-switching approach
• Huiling Zheng (UCL): Fine-grained mortality forecasting with deep learning

14:40 - 15:40 Panel discussion (Room B200)

Effective Collaboration between Academia and Industry in Insurance Data Science (Chair: Dylan Liew)

• Aurélien Couloumy, Dylogy
• Tina Thomson, Gallagher Re
• George Tzougas, Heriot-Watt University
• Malgorzata Wasiewicz, UCL

15:40 - 16:00 Coffee break

16:00 - 17:00 Regular Session 3

Stream 1 Room B200: Explainability (Chair: Rui Zhu)

• Karol Gawlowski, Patricia Wang (EY, Convex): Ensembling GLM and XGB
• Lucas Muzynoski (Avenue Analytics): Fully transparent machine learning: Exact factor table representation of GBMs
• Quentin Guibert, Markéta Krúpová (Université Paris-Dauphine, ADDACTIS): Explainable boosting machine for predicting claim severity and frequency in car insurance

Stream 2 Room BG02: Portfolio Management (Chair: Mick Cooney)

• Francesca Nuzzo, Leonardo Ruggieri (Generali Italia): Strategic asset allocation for insurance product development: A machine learning approach
• Markus Gesmann (Insurance Capital Markets Research): Winning strategies: Predicting relative performance (Slides)
• Freek Holvoet (KU Leuven): Multi-view spatial embeddings for insurance portfolio analytics

17:00 - 18:00 Keynote 2 (Room B200) (Chair: Gráinne McGuire)

James David Long (Palomar): Is Data Engineering the New Data Science?

19:00 Conference dinner

• Ironmongers’ Hall, Shaftesbury Place, Barbican, London EC2Y 8AA

20 June 2025

09:15 - 10:15 Room B200: Keynote 3 (Chair: Mario Wüthrich)

Johanna Ziegel (ETH Zurich): Conformal calibration guarantees for reliable predictions

10:15 - 11:15 Lightning Session 2

Stream 1 Room B200: AI/ML (Chair: Giorgio Alfredo Spedicato)

• Anne van der Scheer (Perunum Actuarieel Advies): From claim counts to interarrival times using a small neural framework
• Yiannis Parizas (Actuary and Open-Source Developer): Advancing non-life insurance modelling with NetSimR
• Michael Ramati (Earnix): Price leakage in demand models
• Gwen Chan, Priyank Shah (Ki Insurance, Lane Clark & Peacock): More data or more model: A framework for achieving predictive analytics objectives in Specialty insurance
• Gustavo Martinez (Mirai Solutions): Agentic AI applications in insurance

Stream 2 Room BG02: Risk Management & Theory (Chair: Davide De March)

• Emilio L. Sáenz Guillén (City St Geoge’s): Non-parametric insurance loss modelling using variable-knot splines
• Brandon Schwab (Leibniz University Hannover): Elevating trust in high-stakes decisions using glass-box models and robust feature selection
• Filip Lindskog (Stockholm University): Claims processing and costs under capacity constraints
• Ran Xu (Xi’an Jiaotong-Liverpool University): LSTM-based coherent mortality forecasting for developing countries
• Giovanni Rabitti (Heriot-Watt University): Analytical variable importance indices for generalized additive models

11:15 - 11:45 Coffee break

11:45 - 12:45 Regular Session 4

Stream 1 Room B200: Regression Methods (Chair: George Tzougas)

• Yuval Ben Dror (Earnix): A two-step regularization algorithm to cluster categories in GLMs
• Mario Wüthrich (ETH Zurich): Tests for auto-calibration
• Tian Dong (UNSW Sydney): Distributional regression for actuarial applications: distributional refinement network

Stream 2 Room BG02: Climate Risks (Chair: Hirbod Assa)

• José Garrido (Concordia University): Catastrophic-risk-aware reinforcement learning with extreme-value-theory-based policy gradients
• Olivier Lopez (Ensae Institut Polytechnique de Paris): Design of parametric insurance via machine learning and optimal combination with traditional insurance
• Mathias Valla, Jose Garrido (France & Aix-Marseille University, Concordia University): Feature and quantile selection for the actuarial climate index: Everything, everywhere, all at once

12:45 - 13:45 Lunch

13:45 - 14:45 Lightning Session 3

Stream 1 Room B200: AI/ML (Chair: Filip Lindskog)

• Tsz Chai Fung (Georgia State University): Statistical learning of trade credit insurance network data with applications to ratemaking and reserving
• Emanuele Fabbiani (Xtream): From SHAP to EBM: How to explain gradient Boosting models
• Meryem Schalck (IPAG Business School): Auto insurance fraud detection: Machine learning and deep learning applications
• Guangyuan Gao (Renmin University of China): Additive tree latent variable models with applications to insurance loss prediction
• Juan Yanez (University of Barcelona): How does granularity affect motor insurance claim predictions in a telematics setting?
• Samuel Gyamerah (Toronto Metropolitan University): Metaheuristic-informed machine learning for optimizing strike temperatures in weather index insurance

Stream 2 Room BG02: Bayes & Graphs (Chair: Despoina Makariou)

• Mark Shoun (Ledger Investing): Copula models of correlation in insurer loss reserves
• Conor Goold (Ledger Investing): Automating tail factor detection points using Bayesian hidden Markov models and latent change-point models
• Paul Wilsens (KU Leuven): Machine learning in an expectation-maximisation framework for nowcasting
• Juan Ignacio de Oyarbide (Addactis): Pricing workers compensation via Bayesian hierarchical modeling
• George Tzougas (Heriot-Watt University): Dynamic hierarchical graph neural networks for spatiotemporal prediction of flood-related claims
• Aurélien Couloumy, Christelle Rovetta (University Lyon 1, Dylogy): Causal knowledge graphs for risk interpretation using LMM: A new tool for insurers

14:45 - 14:55 Room B200: Closing comments (Markus Gesmann)

Gold Sponsors

Posit: We help the world make sense of data.

Ledger Investing: We are building an open insurance system to connect risk to capital

Silver Sponsors

Markel: Bold ideas. Honest actions.

Mirai Solutions: Smarter analytics - better decisions

Ki: Risk, simplified.