Lobster: A GPU-Accelerated Framework for Neurosymbolic Programming

2025-03-27

AI Summary

Existing neural-symbolic learning frameworks suffer from performance bottlenecks and limited scalability due to heterogeneous execution—neural components on GPU and symbolic components on CPU. This paper introduces the first end-to-end GPU-accelerated neural-symbolic programming framework, compiling Datalog-style programs uniformly for GPU execution. Our approach centers on three key contributions: (1) the Abstract Programming Model (APM), an intermediate language enabling joint representation of neural and symbolic logic; (2) a unified abstraction based on justification semirings, supporting differentiable, probabilistic, and discrete reasoning within a single formalism; and (3) a CUDA-optimized compiler that generates efficient GPU-native code. Evaluated across eight cross-domain tasks—including NLP, computer vision, and bioinformatics—the framework achieves a 5.3× average speedup over Scallop and, for the first time, enables real-time inference with large-scale models in these domains.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Knowledge Representation and Reasoning: Logic Programming
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

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 with 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 8 applications, Lobster achieves an average speedup of 5.3x over Scallop, a state-of-the-art neurosymbolic framework, and enables scaling of neurosymbolic solutions to previously infeasible tasks.

Problem

• Bridging GPU-accelerated neural and CPU-based symbolic components
• Enabling end-to-end GPU utilization for neurosymbolic learning
• Improving performance and scalability of neurosymbolic programming frameworks

Innovation

• GPU-accelerated end-to-end neurosymbolic learning framework
• Datalog-based language mapped to GPU programming
• APM intermediate language for flexible reasoning modes