cAIMBER: Causal Artificial Intelligence for Human Mobility Behavior Analysis Using Trajectory Data
The ever-changing mobility landscape and climate change continue to challenge existing operating models and the responsiveness of city planners, policymakers, and regulators. City authorities have growing investment needs that require more focused operations and management strategies that align mobility portfolios to societal goals. The project targets the root cause of traffic (human) and novel analytic techniques to learn and predict human mobility behavior dynamics from pervasive mobile sensing data that can help cities meet both sustainability challenges (through predicting congestion, emissions, and energy consumption) and improve urban resilience to disruptive events (such as infrastructure failures, natural disasters, or pandemics).
The human mobility area witnessed active developments in two broad but separate fields in transport and computer science. They work with different data, use different methods, and answer different but overlapping questions, i.e., mobility behaviour modeling using ‘small’ data in transport and mobility pattern analysis using ‘big’ data in computer science. A solid bridge between these is beneficial and needed but is still an open challenge. Mobile sensing and information technology have enabled us to collect a large amount of mobility trajectory data from human decision-makers. The predictive AI techniques show the potential to efficiently learn and predict human mobility using these trajectory data. However, they continually run up against the limits of what they observe (correlations, not causal relationships), thus hindering any type of serious applicability for preparedness and response policies for cities without understanding the causal mobility dynamics.
The project aims to develop scalable and causal AI methodologies to analyze and predict human mobility behavior dynamics using individual travel trajectory data. Also, importantly the project will use large-scale, longitudinal smartcard metadata to develop causal diagrams of human mobility behavior under disturbances that could help design effective strategies for sustainable and resilient urban mobility systems. The key idea is to formulate the mobility behavior dynamics learning task as a causal reinforcement learning problem, which combines the RL method (efficient in learning) and the causal inference model (reasoning about counterfactual nature). The algorithm learns the structural causal environment model, conditional policy and reward functions, and casual diagram using observational and interventional data.
PI: Zhenliang Ma, Transportation Science. Co-PI; Liam Solus, Mathematics