Haedar Ahmed Mukhef

Memory-Efficient Probabilistic Neuro-Symbolic Integration for Explainable Natural Language Inference Using Transformer-Based Foundation Models

A memory-optimized probabilistic neuro-symbolic hybrid architecture that unifies transformer-based neural networks with logic-based symbolic reasoning systems within memory-constrained systems indicates that neuro-symbolic systems within memory-constrained systems can achieve both explanation needs and foundation models’ performance requirements.