Andreas Prein (ETH Zurich)
Abstract:
Convective storms are responsible for substantial societal and economic losses around the world. These impacts are rising - primarily due to increasing exposure - yet the underlying hazard remains deeply uncertain, even in regions with dense observational networks. This uncertainty complicates efforts to anticipate and manage risk. In this keynote, we will present recent advances aimed at enhancing our understanding of global convective storm hazards through the use of both statistical and dynamic modeling approaches. Statistical models, including machine learning techniques, are increasingly used to estimate hail hazard at continental to global scales, providing valuable insights into large-scale patterns and regional variability. Complementing these efforts, convection-permitting simulations - run at kilometer-scale resolution on regional and global scales - allow for the explicit representation of severe convective processes. These include extreme precipitation, hail, straight-line winds, and supercell thunderstorms. We will discuss the strengths and weaknesses of these approaches and highlight how combining them provides a more comprehensive understanding of convective storm hazards. We will conclude by highlighting open questions, future research directions, and opportunities for public-private partnerships, particularly in consideration of ongoing and projected climate changes, to support improved hazard assessment and risk reduction strategies globally.
14:00 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room E 7
Marco Bacci (Blue Marble Micro Insurance)
Abstract:
This presentation explores how parametric insurance, powered by remote sensing and global partnerships, can enable risk protection for underserved populations. We introduce the fundamentals of parametric insurance and Blue Marble, a company delivering scalable solutions for low-income, climate-vulnerable communities. The talk demonstrates how remote sensing enables fast, transparent insurance through Blue Marble’s weather-based indices in contexts characterized by low income and limited financial literacy. Ultimately, we show how balancing simplicity and innovation can strengthen resilience and support climate adaptation for vulnerable communities.
14:35 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room E 7
Ruedi Bodenmann (Assura)
Abstract:
Quantitative risk modeling of financial impact on earnings and capital is essential in insurance. But risk management goes way beyond risk modeling. In order to protect P&L and balance sheet as well as customer relationships and operating resources from adverse scenarios and to guarantee the realization of strategic targets, an adequate risk governance and a risk culture need to be set up. Corrective actions have to be put in place which systematically respect the criteria of adequacy and completeness, effectiveness, and efficiency. Real life examples will be discussed.
15:35 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room E 7
Marc Beierschoder and Reto Haeni (Deloitte)
Abstract:
AI is shifting from pilots to enterprise scale, creating new opportunities and risks. In this session, we will share lessons from current projects and explore the rise of Agentic AI systems that act autonomously rather than just assist. We will introduce Deloitte’s Trustworthy AI Framework and show how it helps organizations embed governance, ethics, and accountability while accelerating real impact.
16:10 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room E 7
Uni ???
Adrien Hardy (Qube Research & Technologies)
Abstract:
Data augmentation refers to the modern practice of enriching a dataset with generated samples, with the aim of improving statistical model estimation. While it has proven highly effective in structured data contexts such as image or text analysis, its benefits in the domain of financial time series remain unclear - despite a growing number of empirical studies. In this joint work with Lucas Morisset (QRT) and Alain Durmus (École Polytechnique), we quantify the impact of a data generation scheme on a canonical task in the field: inverse covariance matrix estimation. This problem is closely related to multivariate linear regression parameter estimation and arises in applications such as optimal portfolio allocation under a mean–variance tradeoff. Our results rely on new deterministic equivalents for generalized resolvents of random covariance matrices, which may be of independent interest.
17:10 • Location T.B.A.
Wenqian Huang (Bank for International Settlements)
Abstract:
This talk explores how decentralized finance (DeFi) is introducing new mechanisms for trading, lending, and even investing in real-world assets, while highlighting the trade-offs that come with these innovations. It examine decentralized exchanges such as Uniswap and Curve, where pooled liquidity and automated market-making reduce the price impact of large trades – for example, the launch of Curve’s stablecoin exchange significantly dampened volatility in stablecoin trading pairs. We will also explain how DeFi lending platforms like Aave allow crypto-collateralized loans with on-chain, automatic liquidation triggers that protect lenders but can quickly close positions when collateral values drop. In addition, we discuss the tokenization of illiquid assets like real estate, demonstrating how assets can be split into tradable tokens to increase market access. However, evidence shows that leverage-fueled buying in tokenized real estate can drive up prices and volatility, underscoring the importance of understanding risk. By connecting insights from recent research to real-world examples (Aave, Uniswap, Curve, RealT), the talk illustrates that while DeFi offers novel ways to trade and lend, its economic trade-offs must be understood to maintain market quality and financial stability.
17:45 • ETH Zentrum, Rämistrasse 101, Zürich, Building HG, Room E 7