analiza aiqbench
This paper addresses the scalability bottleneck in Neurosymbolic AI by optimizing the NeurASP framework with vectorization, batch processing, and caching, achieving significant speedups for complex tasks. The approach effectively mitigates the high computational cost of backpropagation through non-differentiable symbolic components, enabling the training of more sophisticated models. A new dataset featuring difficult playing card tasks is introduced to validate the enhanced learning capabilities of the optimized system.
Accelerating NeurASP with vectorization and caching
arXiv:2606.10787v1 Announce Type: new Abstract: Neurosymbolic AI combines neural networks with symbolic programs to create robust and explainable predictions. One such framework is NeurASP, which trains a neural network to predict concepts and reasons over them using rules written in answer set programming (ASP) to solve downstream tasks. Crucially, labels are only provided for the downstream prediction produced by the symbolic rules, not for the latent concepts themselves.Backpropagation through the non-differentiable ASP component requires expensive probability and gradient calculations, which has hindered scalability to more sophisticated tasks.In this paper, we address the current limitations of NeurASP by improving its computational performance through vectorization, batch processing and caching of intermediate computations during training. We compare computation speeds between the original and our new implementation of NeurASP and report speedups of multiple orders of magnitude for larger tasks. To this end, we propose a new dataset of difficult tasks involving playing cards, which we use to test the capabilities of NeurASP's enhanced learning function.
Abstract:Neurosymbolic AI combines neural networks with symbolic programs to create robust and explainable predictions. One such framework is NeurASP, which trains a neural network to predict concepts and reasons over them using rules written in answer set programming (ASP) to solve downstream tasks. Crucially, labels are only provided for the downstream prediction produced by the symbolic rules, not for the latent concepts this http URL through the non-differentiable ASP component requires expensive probability and gradient calculations, which has hindered scalability to more sophisticated this http URL this paper, we address the current limitations of NeurASP by improving its computational performance through vectorization, batch processing and caching of intermediate computations during training. We compare computation speeds between the original and our new implementation of NeurASP and report speedups of multiple orders of magnitude for larger tasks. To this end, we propose a new dataset of difficult tasks involving playing cards, which we use to test the capabilities of NeurASP's enhanced learning function.
Artificial Intelligence (cs.AI); Logic in Computer Science (cs.LO)
arXiv:2606.10787 [cs.AI]
(or arXiv:2606.10787v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2606.10787 arXiv-issued DOI via DataCite (pending registration)
Text integral sub licența deschisă a publicației · 184 cuvinte
Subiecte conexe
Accelerate Vector Diffusion Maps by Landmarks
arXiv · 1 sept. 2026
Multi-Robot Motions in Milliseconds: Vector-Accelerated Primitives for Sampling-Based Planning
arXiv · 7 aug. 2026
From Vector Search to Agentic RAG: Building an Enterprise Research Analyst with hipVS
AMD · 15 iul. 2026
D-NOVA: In-Storage Retrieval Accelerator via Dual-Bound 3D NAND-Optimized Similarity Search with Vector Adaptation
arXiv · 21 iul. 2026