Causal knowledge extraction
Developed and evaluated methods for extracting statements about causal relationships from text and organizing them for event analysis.
Causal Models & Enterprise Risk / 2018–2024
Engine for Event Forecasting and Scenario Planning
Forecasting events and exploring risk scenarios.
I led the Engine for Forecasting Significant Societal Events research program. We studied causal knowledge extraction and probabilistic event models, and developed forecasting tools. This work contributed to IBM’s Scenario Planning Advisor, deployed for the IBM Chief Risk Office and recognized with the ICAPS 2020 Best System Demonstration Award.
Graphical and probabilistic models
Studied event sequences using Summary Markov Models, rule-based extensions, and neural prediction methods.
Enterprise scenario planning
Contributed forecasting research to a deployed enterprise risk system that combines learned models with planning-based hypothesis generation.