NeSyPr: Neurosymbolic Proceduralization For Efficient Embodied Reasoning
This work addresses efficiency and deployment issues for LM-based agents in resource-constrained physical systems, representing an incremental improvement by combining neurosymbolic methods for specific bottlenecks.
The paper tackles the challenge of using language models for embodied tasks in dynamic environments with latency and resource constraints by introducing NeSyPr, a neurosymbolic proceduralization framework that compiles symbolic plans into procedural representations for efficient LM inference, demonstrating improved reasoning capabilities on benchmarks like PDDLGym, VirtualHome, and ALFWorld with more compact models.
We address the challenge of adopting language models (LMs) for embodied tasks in dynamic environments, where online access to large-scale inference engines or symbolic planners is constrained due to latency, connectivity, and resource limitations. To this end, we present NeSyPr, a novel embodied reasoning framework that compiles knowledge via neurosymbolic proceduralization, thereby equipping LM-based agents with structured, adaptive, and timely reasoning capabilities. In NeSyPr, task-specific plans are first explicitly generated by a symbolic tool leveraging its declarative knowledge. These plans are then transformed into composable procedural representations that encode the plans' implicit production rules, enabling the resulting composed procedures to be seamlessly integrated into the LM's inference process. This neurosymbolic proceduralization abstracts and generalizes multi-step symbolic structured path-finding and reasoning into single-step LM inference, akin to human knowledge compilation. It supports efficient test-time inference without relying on external symbolic guidance, making it well suited for deployment in latency-sensitive and resource-constrained physical systems. We evaluate NeSyPr on the embodied benchmarks PDDLGym, VirtualHome, and ALFWorld, demonstrating its efficient reasoning capabilities over large-scale reasoning models and a symbolic planner, while using more compact LMs.