arXiv subjects

Mahmood Fazlali

Publications and source records attributed to Mahmood Fazlali.

3 recordsLinked to original sources

arXiv2026

Forward-Only Convolutional Neural Networks with Learnable Channel-Class Assignment

The Forward-Forward (FF) algorithm offers a biologically inspired alternative to backpropagation by replacing gradient-based credit assignment with local, forward-only objectives. While recent extensions have adapted FF to convolutional neural networks (CNNs), existing formulations rely on static channel-class partitions and struggle to perform effectively in complex tasks. In this work, we introduce a learnable channel-class assignment mechanism that enables adaptive, data-driven specialization of convolutional channels, supported by entropy and orthogonality regularization to promote learning performance. We further propose a loss-aware layer contribution strategy that adaptively weights intermediate-layer predictions based on their validation performance, enhancing the effectiveness of forward-only inference. Integrated into residual CNNs, the proposed method achieves consistently superior performance across CIFAR-10, CIFAR-100, and Tiny-ImageNet compared to existing similar forward-only methods. Notably, it establishes new state-of-the-art performance among FF-based models, substantially narrowing the gap with backpropagation. These findings demonstrate that introducing learnable channel specialization and layer contribution weighting significantly enhances the representational capacity of forward-only learning in deep CNNs.

cs.LG↗

arXiv2025

Backpropagation-free Spiking Neural Networks with the Forward-Forward Algorithm

Spiking Neural Networks (SNNs) offer a biologically inspired computational paradigm that emulates neuronal activity through discrete spike-based processing. Despite their advantages, training SNNs with traditional backpropagation (BP) remains challenging due to computational inefficiencies and a lack of biological plausibility. This study explores the Forward-Forward (FF) algorithm as an alternative learning framework for SNNs. Unlike backpropagation, which relies on forward and backward passes, the FF algorithm employs two forward passes, enabling layer-wise localized learning, enhanced computational efficiency, and improved compatibility with neuromorphic hardware. We introduce an FF-based SNN training framework and evaluate its performance across both non-spiking (MNIST, Fashion-MNIST, Kuzushiji-MNIST) and spiking (Neuro-MNIST, SHD) datasets. Experimental results demonstrate that our model surpasses existing FF-based SNNs on evaluated static datasets with a much lighter architecture while achieving accuracy comparable to state-of-the-art backpropagation-trained SNNs. On more complex spiking tasks such as SHD, our approach outperforms other SNN models and remains competitive with leading backpropagation-trained SNNs. These findings highlight the FF algorithm's potential to advance SNN training methodologies by addressing some key limitations of backpropagation.

cs.NE↗

arXiv2018

On complexity of post-processing in analyzing GATE-driven X-ray spectrum

Computed Tomography (CT) imaging is one of the most influential diagnostic methods. In clinical reconstruction, an effective energy is used instead of total X-ray spectrum. This approximation causes an accuracy decline. To increase the contrast, single source or dual source dual energy CT can be used to reach optimal values of tissue differentiation. However, these infrastructures are still at the laboratory level, and their safeties for patients are still yet to mature. Therefore, computer modelling of DECT could be used. We propose a novel post-processing approach for converting a total X-ray spectrum into irregular intervals of quantized energy. We simulate a phantom in GATE/GEANT4 and irradiate it based on CT configuration. Inverse Radon transform is applied to the acquired sinogram to construct the Pixel-based Attenuation Matrix (PAM). To construct images represented by each interval, water attenuation coefficient of the interval is extracted from NIST and used in the Hounsfield unit (HU) scale in conjunction with PAM. The CT image is modified by using of an associated normalized photon flux and calculated HU corresponding to the interval. We demonstrate the proposed method efficiency via complexity analysis, using absolute and relative complexities, entropy measures, Kolmogorov complexity, morphological richness, and quantitative segmentation criteria associated with standard fuzzy C-means. The irregularity of the modified CT images decreases over the simulated ones.

cs.ET↗