Answers are grounded in this paper and cite their sources.
This work reports a simplified, biologically inspired predictive local learning rule that eliminates the need for global backpropagation in conventional neural networks and membrane integration in neuromorphic computing. Weight updates are triggered only on prediction errors and are performed using sparse, binary-driven vector additions. We integrate this rule into an extreme learning machine (ELM), replacing the conventional computationally intensive matrix inversion. Compared to standard ELM, our approach reduces training complexity from to in terms of hidden-layer nodes, while maintaining comparable accuracy averaged across different training and test datasets. We validate performance on various open-source and in-house-collected datasets under strong noise and class imbalance. We also demonstrate a field-programmable gate array (FPGA) implementation and compare it with prior studies, showing significant reductions in computational and memory requirements and higher throughput. This design demonstrates strong potential for energy-efficient online learning on low-cost edge devices.
No takes yet. Share an insight, caveat, or question.
Zang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: