Lobster: A GPU-Accelerated Framework for Neurosymbolic Programming

TL;DR

Lobster addresses the bottleneck of CPU-bound symbolic reasoning in neurosymbolic programming by mapping a Datalog-based language onto GPUs through a new intermediate language, APM. It enables end-to-end GPU execution for discrete, probabilistic, and differentiable reasoning via a provenance semiring framework and a GPU-oriented compiler/runtime implemented in Rust and CUDA. The approach yields substantial performance gains, including an average speedup of about 3.9x over Scallop across nine applications and enabling tasks previously infeasible due to symbolic bottlenecks. The work demonstrates broad applicability across NLP, image processing, program analysis, bioinformatics, and planning, and provides practical optimizations, a scalable architecture, and open-source tooling for future neurosymbolic systems on GPUs.

Abstract

Neurosymbolic programs combine deep learning with symbolic reasoning to achieve better data efficiency, interpretability, and generalizability compared to standalone deep learning approaches. However, existing neurosymbolic learning frameworks implement an uneasy marriage between a highly scalable, GPU-accelerated neural component and a slower symbolic component that runs on CPUs. We propose Lobster, a unified framework for harnessing GPUs in an end-to-end manner for neurosymbolic learning. Lobster maps a general neurosymbolic language based on Datalog to the GPU programming paradigm. This mapping is implemented via compilation to a new intermediate language called APM. The extra abstraction provided by apm allows Lobster to be both flexible, supporting discrete, probabilistic, and differentiable modes of reasoning on GPU hardware with a library of provenance semirings, and performant, implementing new optimization passes. We demonstrate that Lobster programs can solve interesting problems spanning the domains of natural language processing, image processing, program reasoning, bioinformatics, and planning. On a suite of 9 applications, Lobster achieves an average speedup of 3.9x over Scallop, a state-of-the-art neurosymbolic framework, and enables scaling of neurosymbolic solutions to previously infeasible tasks.

Figures (14)

• Figure 1: An example neurosymbolic program.
• Figure 2: High-level comparison between Lobster and existing frameworks.
• Figure 3: The overall pipeline, symbolic program, and the acceleration result of using Lobster for the Pathfinder task.
• Figure 4: The RAM language.
• Figure 5: Provenance semiring structure and examples.