Meet CompatibL at the 22nd Quantitative Finance Conference

Aug 25, 2026

We are excited to announce that CompatibL will exhibit and present at the 22nd Quantitative Finance Conference, organized by World Business Strategies (WBS).

The event will take place in the historic city of Valletta, Malta, from September 30 through October 2, 2026, at the Grand Hotel Excelsior.

This year, CompatibL’s Executive Chairman and Head of Quant Research, Alexander Sokol, will present two cutting-edge research topics, lead a dedicated workshop, and moderate a panel discussion.

Workshop: AI Model Risk Measurement, Reporting, and Mitigation

Wednesday, September 30, 13:45–17:15 CEST

Established model risk management practices often fall short when faced with the unique characteristics of AI models. In this workshop, Alexander will present practical techniques for the quantitative measurement and reporting of AI model risk using both established and novel metrics, and will describe how to mitigate AI model risk and increase the reliability of AI-based workflows.

Panel Discussion: Is Active Deception Part of AI Model Risk?

Thursday, October 1, 9:45–10:30 CEST

This panel will explore active deception by AI models through the lens of adversarial risk analysis (ARA). Key topics include test awareness (how models alter behavior when evaluated), intentional noncompliance, techniques for measuring AI model risk in the presence of active deception, and the use of cyber defense strategies to mitigate these risks.

Presentation 1: Adversarial Risk Analysis of AI Models

Thursday, October 1 (AI / LLMs / ML Stream), 11:45–12:30 CEST

Alexander will discuss how AI models, like humans, can engage in active deception and strategic reasoning when they detect evaluation controls. He will talk about Adversarial Risk Analysis (ARA) as a quantitative framework designed to assess and manage active deception during model testing—addressing capabilities that evade standard AI risk controls.

Presentation 2: Currency Pooling for Estimating Curve Basis Representations and Mean Reversion Rates

Thursday, October 1 (Trading & Interest Rate Modeling Stream), 15:45–16:30 CEST

In this session, Alexander will present statistical evidence of a universal interest rate manifold across nine major currencies, where data excluding a currency predicts its future rates better than its own data. Building on these findings, he will propose a technique for estimating curve basis representations and mean-reversion rates by pooling data from multiple currencies while retaining standard currency-specific volatility calibrations.

About WBS’s Quantitative Finance Conference

Every year, WBS’s Quantitative Finance Conference brings together quantitative finance professionals, researchers, academics, and industry experts from around the world. The event provides a collaborative platform to exchange ideas, showcase groundbreaking research, and discuss emerging trends in financial technology and quantitative modeling.