CASSANDRA: Programmatic and Probabilistic Learning and Inference for Stochastic World Modeling
This addresses the challenge of modeling complex action effects and causal relationships from limited data for planning in business domains, representing an incremental advancement by integrating LLMs with probabilistic methods.
The paper tackled the problem of building world models for planning in real-world domains by proposing CASSANDRA, a neurosymbolic approach that uses an LLM as a knowledge prior to construct lightweight transition models, resulting in significant improvements in transition prediction and planning over baselines in simulators like a coffee shop and theme park business.
Building world models is essential for planning in real-world domains such as businesses. Since such domains have rich semantics, we can leverage world knowledge to effectively model complex action effects and causal relationships from limited data. In this work, we propose CASSANDRA, a neurosymbolic world modeling approach that leverages an LLM as a knowledge prior to construct lightweight transition models for planning. CASSANDRA integrates two components: (1) LLM-synthesized code to model deterministic features, and (2) LLM-guided structure learning of a probabilistic graphical model to capture causal relationships among stochastic variables. We evaluate CASSANDRA in (i) a small-scale coffee-shop simulator and (ii) a complex theme park business simulator, where we demonstrate significant improvements in transition prediction and planning over baselines.