Yongzhao Wang - Research Areas

Automated Strategic Analysis from Natural Language

A primary challenge in conducting game-theoretic analysis from natural language descriptions of complex scenarios is the initial construction of a precise formal representation of the underlying game. In game description translation, large language models are used to derive such representations in forms that are directly compatible with established game-theoretic methods, thereby enabling automated game-theoretic analysis from natural-language inputs.

Empirical Game-Theoretic Analysis

The methodology of Empirical Game-Theoretic Analysis (EGTA) provides a comprehensive suite of techniques for strategic reasoning in games using models derived from simulation data. For multiagent systems not amenable to analytic solution, EGTA provides a simulation-based alternative, addressing the most important strategic considerations.

Computational Finance

In computational finance, we are concerned with applying AI to address practical, real-world problems in modern financial systems, from algorithmic trading in asset markets to the enhancement of financial services.

Adversarial Reasoning in Cybersecurity

In cybersecurity, interactions between attackers and defenders can be modeled as a strategic contest in which each side adapts to the other under uncertainty, limited resources, and imperfect information. Game theory offers a formal framework for modeling and analyzing these dynamics, and, when combined with AI, can scale to large and complex scenarios.