Building Intelligent Systems for Finance, Insurance, and Decentralized Markets

As financial systems undergo rapid transformation, the convergence of decentralized technologies, AI, and knowledge representation presents new possibilities and new challenges for ensuring trust, transparency, and resilience across diverse domains, from decentralized finance (DeFi) protocols to insurance and over-the-counter (OTC) derivatives contracts. This talk introduces a vision for building intelligent, privacy-preserving, and interoperable systems that empower stakeholders while mitigating systemic risks. In this talk, we will highlight recent advances in secure and explainable federated learning for financial crime detection, designed to protect user privacy while detecting anomalies across siloed datasets. We will also examine the emerging field of Large Transaction Models (LTMs), i.e., foundation models fine-tuned on financial transaction data to power dynamic use cases such as customer segmentation, forecasting under "what-if" scenarios, and fraud prediction. By combining decentralized infrastructure with vertical and hybrid federated learning, credit allocation incentives, and semantic knowledge representation, we showcase how resilient, accountable, and intelligent networked systems can redefine how trust is earned, managed, and operationalized in financial, insurance, and decentralized ecosystems.