Ahmed Louri
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Ahmed Louri has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Ke Wang
University of North Carolina at Charlotte, Charlotte, NC, USA
10 shared publications - Hao Zheng
University of Central Florida, Orlando, FL, USA
8 shared publications - Avinash Karanth
Ohio University
7 shared publications - Fabrizio Lombardi
Northeastern University
6 shared publications - Shanshan Liu
University of Electronic Science and Technology of China
6 shared publications - Yingnan Zhao
George Washington University, Washington, DC, USA
5 shared publications
Latest Publications
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 25 (2) , 440-443
Cross-Layer Head Prediction and KV-Cache Prefetching for High-Performance MoH LLM
- 1 July 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Computer Architecture Letters
- Vol. 25 (2) , 440-443
Cross-Layer Head Prediction and KV-Cache Prefetching for High-Performance MoH LLM
- Open Access
- 22 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 367-373
Learning-Assisted Adaptive Hybrid Interconnection Design for Chiplet-Based Heterogeneous Systems
- 22 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 367-373
Open AccessLearning-Assisted Adaptive Hybrid Interconnection Design for Chiplet-Based Heterogeneous Systems
- 20 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 37 (6) , 1430-1441
Bal-DGCN: A Hardware Acceleration Framework for Balanced Computational Efficiency in DGCNs
Dynamic Graph Convolutional Networks (DGCNs) have emerged as a powerful approach for analyzing dynamic graph-structured data across various applications, such as social networks, recommendation systems, and other domains. Typically, each DGCN layer comprises two distinct modules: a Graph Convolutional Network (GCN) module and a Recurrent Neural Network (RNN) module with unique data communication and computation patterns. Although several customized DGCN accelerators have achieved significant improvements, they primarily focus on either enhancing computational efficiency through specialized Processing Element (PE) designs or reducing redundant memory access via algorithm-level optimizations. Many have not exploited data reuse, which is becoming increasingly critical as the size of input dynamic graphs grows. This growth leads to larger and denser intermediate matrices during matrix multiplications, resulting in a substantial increase in memory access. Additionally, current DGCN accelerators ignore the sparsity during inference processing, leading to uneven workload distribution across the hardware platform and resulting in hardware underutilization. The excessive memory access and hardware underutilization ultimately degrade performance and energy efficiency. To address these challenges, this paper proposes Bal-DGCN, a DGCN hardware accelerator that consists of three key innovations: a unified PE array, a versatile interconnection fabric, and a dynamic control policy. Specifically, Bal-DGCN implements a unified PE array to efficiently perform key computations across the GCN and RNN modules, improving computational efficiency. To manage data communication among PEs, Bal-DGCN integrates a versatile interconnection fabric that supports various on-chip data communication patterns, thereby improving data reuse efficiency. Additionally, Bal-DGCN introduces a dynamic control policy that selectively allocates computational workloads to specific hardware resources to mitigate workload imbalance, maximizing hardware utilization. By significantly improving computational efficiency, data reuse, and hardware utilization, Bal-DGCN achieves substantial gains in both performance and energy efficiency. Evaluation results show that Bal-DGCN achieves an average reduction of 76% in execution time and 72% in energy consumption for DGCN inference compared to existing approaches.
- 20 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Parallel and Distributed Systems
- Vol. 37 (6) , 1430-1441
Bal-DGCN: A Hardware Acceleration Framework for Balanced Computational Efficiency in DGCNs
Dynamic Graph Convolutional Networks (DGCNs) have emerged as a powerful approach for analyzing dynamic graph-structured data across various applications, such as social networks, recommendation systems, and other domains. Typically, each DGCN layer comprises two distinct modules: a Graph Convolutional Network (GCN) module and a Recurrent Neural Network (RNN) module with unique data communication and computation patterns. Although several customized DGCN accelerators have achieved significant improvements, they primarily focus on either enhancing computational efficiency through specialized Processing Element (PE) designs or reducing redundant memory access via algorithm-level optimizations. Many have not exploited data reuse, which is becoming increasingly critical as the size of input dynamic graphs grows. This growth leads to larger and denser intermediate matrices during matrix multiplications, resulting in a substantial increase in memory access. Additionally, current DGCN accelerators ignore the sparsity during inference processing, leading to uneven workload distribution across the hardware platform and resulting in hardware underutilization. The excessive memory access and hardware underutilization ultimately degrade performance and energy efficiency. To address these challenges, this paper proposes Bal-DGCN, a DGCN hardware accelerator that consists of three key innovations: a unified PE array, a versatile interconnection fabric, and a dynamic control policy. Specifically, Bal-DGCN implements a unified PE array to efficiently perform key computations across the GCN and RNN modules, improving computational efficiency. To manage data communication among PEs, Bal-DGCN integrates a versatile interconnection fabric that supports various on-chip data communication patterns, thereby improving data reuse efficiency. Additionally, Bal-DGCN introduces a dynamic control policy that selectively allocates computational workloads to specific hardware resources to mitigate workload imbalance, maximizing hardware utilization. By significantly improving computational efficiency, data reuse, and hardware utilization, Bal-DGCN achieves substantial gains in both performance and energy efficiency. Evaluation results show that Bal-DGCN achieves an average reduction of 76% in execution time and 72% in energy consumption for DGCN inference compared to existing approaches.
- Open Access
- 26 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 24 (5s) , 1-26
FORT-GCN: A Fault-tolerant and Adaptive Accelerator Design for Efficient Graph Convolutional Network Inference
Hardware reliability has emerged as a paramount concern for machine learning accelerators, as transient errors and permanent failures occurring during inference can severely compromise accuracy, performance, and service availability. Although fault resilience in traditional machine learning, such as Deep Neural Networks (DNNs), has been extensively studied, graph convolutional networks (GCNs) present unique reliability challenges due to their irregular computation patterns and dynamic data dependencies. Traditional fault mitigation approaches, including hardware redundancy, recomputation, and Hamming code protection, suffer from prohibitive latency and power overheads when applied to GCN accelerators. This paper presents FORT-GCN, a holistic hardware architecture co-optimized for GCN-specific fault resilience. Our solution integrates three key innovations, namely permanent fault tolerance through a novel robust processing element design with runtime reconfiguration and defect-adaptive interconnects, transient error resilience via lightweight selective error correction unit design, and a fault-aware adaptive controller design that dynamically adjusts fault protection strategies based on operational faults and graph characteristics. Experimental evaluation demonstrates 35.4% improvement in fault robustness compared to conventional error-correction and redundancy-based approaches, with minimal timing, area, and power overheads.
- 26 September 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Embedded Computing Systems
- Vol. 24 (5s) , 1-26
Open AccessFORT-GCN: A Fault-tolerant and Adaptive Accelerator Design for Efficient Graph Convolutional Network Inference
Hardware reliability has emerged as a paramount concern for machine learning accelerators, as transient errors and permanent failures occurring during inference can severely compromise accuracy, performance, and service availability. Although fault resilience in traditional machine learning, such as Deep Neural Networks (DNNs), has been extensively studied, graph convolutional networks (GCNs) present unique reliability challenges due to their irregular computation patterns and dynamic data dependencies. Traditional fault mitigation approaches, including hardware redundancy, recomputation, and Hamming code protection, suffer from prohibitive latency and power overheads when applied to GCN accelerators. This paper presents FORT-GCN, a holistic hardware architecture co-optimized for GCN-specific fault resilience. Our solution integrates three key innovations, namely permanent fault tolerance through a novel robust processing element design with runtime reconfiguration and defect-adaptive interconnects, transient error resilience via lightweight selective error correction unit design, and a fault-aware adaptive controller design that dynamically adjusts fault protection strategies based on operational faults and graph characteristics. Experimental evaluation demonstrates 35.4% improvement in fault robustness compared to conventional error-correction and redundancy-based approaches, with minimal timing, area, and power overheads.
- 23 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 15 (4) , 560-576
Extending Energy-Efficient and Scalable DNN Training and Inference With 3-D Photonic Accelerator
As deep neural network (DNN) models continue to grow in complexity, analog computing architectures have emerged as a promising solution to meet increasing computational demands. Among these, silicon photonic computing excels at efficiently executing dot product operations while leveraging inherent parallelism. Photonic phase change memory (photonic-PCM) further enhances photonic computing by enabling scalable, non-volatile storage. In this work, we introduce the 3D Large-Scale Photonic Accelerator (LSPA), a novel photonic computing architecture designed for large-scale DNN models. LSPA employs multi-layered 3D stacking of non-volatile photonic-PCM cells, creating a high-density computational fabric that optimizes energy efficiency, flexibility, and scalability. LSPAs custom 3D photonic network enables simultaneous data multicast in two dimensions and accumulation in three dimensions, optimizing communication patterns essential for efficient DNN training. A distinctive feature of LSPA is its ability to execute multiple forward and backward passes in parallel within each mini-batch, reducing latency associated with data movement and photonic-PCM programming. This unique capability combined with high-bandwidth photonic interconnects allows LSPA to sustain efficient training across a wide range of DNN workloads. When evaluated against a range of neural network models including VGG-16, ResNet-50, GoogLeNet, Transformer, GNMT, LLaMA 7B, and LLaMA 30B, LSPA reduces execution time by up to 92% and energy consumption by up to 90%. These results highlight LSPA as a transformative advancement in scalable, high-performance photonic computing for deep learning.
- 23 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Journal on Emerging and Selected Topics in Circuits and Systems
- Vol. 15 (4) , 560-576
Extending Energy-Efficient and Scalable DNN Training and Inference With 3-D Photonic Accelerator
As deep neural network (DNN) models continue to grow in complexity, analog computing architectures have emerged as a promising solution to meet increasing computational demands. Among these, silicon photonic computing excels at efficiently executing dot product operations while leveraging inherent parallelism. Photonic phase change memory (photonic-PCM) further enhances photonic computing by enabling scalable, non-volatile storage. In this work, we introduce the 3D Large-Scale Photonic Accelerator (LSPA), a novel photonic computing architecture designed for large-scale DNN models. LSPA employs multi-layered 3D stacking of non-volatile photonic-PCM cells, creating a high-density computational fabric that optimizes energy efficiency, flexibility, and scalability. LSPAs custom 3D photonic network enables simultaneous data multicast in two dimensions and accumulation in three dimensions, optimizing communication patterns essential for efficient DNN training. A distinctive feature of LSPA is its ability to execute multiple forward and backward passes in parallel within each mini-batch, reducing latency associated with data movement and photonic-PCM programming. This unique capability combined with high-bandwidth photonic interconnects allows LSPA to sustain efficient training across a wide range of DNN workloads. When evaluated against a range of neural network models including VGG-16, ResNet-50, GoogLeNet, Transformer, GNMT, LLaMA 7B, and LLaMA 30B, LSPA reduces execution time by up to 92% and energy consumption by up to 90%. These results highlight LSPA as a transformative advancement in scalable, high-performance photonic computing for deep learning.
- Open Access
- 20 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1240-1253
DiTile-DGNN: An Efficient Accelerator for Distributed Dynamic Graph Neural Network Inference
- 20 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 1240-1253
Open AccessDiTile-DGNN: An Efficient Accelerator for Distributed Dynamic Graph Neural Network Inference
- 25 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Sustainable Computing
- Vol. 10 (4) , 705-716
MERIT: A Sustainable DNN Accelerator Design With Photonic Phase-Change Memory
The growing computational demands of deep learning have driven interest in analog neural networks using resistive memory and silicon photonics. However, these technologies face inherent limitations in computing parallelism when used independently. Photonic phase-change memory (PCM), which integrates photonics with PCM, overcomes these constraints by enabling simultaneous processing of multiple inputs encoded on different wavelengths, significantly enhancing parallel computation for deep neural network (DNN) inference and training. This paper presents MERIT, a sustainable DNN accelerator that capitalizes on the non-volatility of resistive memory and the high operating speed of photonic devices. MERIT enables seamless inference and training by loading weight kernels into photonic PCM arrays and selectively supplying light encoded with input features for the forward pass and loss gradients for the backward pass. We compare MERIT with state-of-the-art digital and analog DNN accelerators including TPU, DEAP, and PTC. Simulation results demonstrate that MERIT reduces execution time by 68% and energy consumption by 64% for inference, and reduces execution time by 79% and energy consumption by 84% for training.
- 25 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Sustainable Computing
- Vol. 10 (4) , 705-716
MERIT: A Sustainable DNN Accelerator Design With Photonic Phase-Change Memory
The growing computational demands of deep learning have driven interest in analog neural networks using resistive memory and silicon photonics. However, these technologies face inherent limitations in computing parallelism when used independently. Photonic phase-change memory (PCM), which integrates photonics with PCM, overcomes these constraints by enabling simultaneous processing of multiple inputs encoded on different wavelengths, significantly enhancing parallel computation for deep neural network (DNN) inference and training. This paper presents MERIT, a sustainable DNN accelerator that capitalizes on the non-volatility of resistive memory and the high operating speed of photonic devices. MERIT enables seamless inference and training by loading weight kernels into photonic PCM arrays and selectively supplying light encoded with input features for the forward pass and loss gradients for the backward pass. We compare MERIT with state-of-the-art digital and analog DNN accelerators including TPU, DEAP, and PTC. Simulation results demonstrate that MERIT reduces execution time by 68% and energy consumption by 64% for inference, and reduces execution time by 79% and energy consumption by 84% for training.
- 22 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 13 (3) , 714-723
Fault Tolerance in Triplet Network Training: Analysis, Evaluation and Protection Methods
This paper investigates the tolerance of Triplet Networks (TNs) with a focus on faults in the training process. For compatibility with the existing literature. So-called stuck-at faults of a functional nature are considered for the operation of the neurons and activation function. While TNs are shown to be generally robust against such faults in the anchor and positive subnetworks, the presented analysis reveals a significant vulnerability in the negative subnetwork, in which stuck-at faults can lead to false convergence and training failures. An in-depth treatment is provided to show the incorrect convergence of training in the presence of stuck-at faults, highlighting the behavior of the network with faulty neurons. Extensive simulations are presented to evaluate the impact of these faults and propose two innovative fault-tolerant methods: the regularization of the anchor outputs and the modified margin. Simulation shows that false convergence can be very efficiently avoided by utilizing the proposed techniques, and thus the overall accuracy loss of the TN is negligible. These findings contribute to the understanding of fault tolerance in emerging neural networks such as TNs and offer practical solutions for enhancing their robustness against faults.
- 22 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Emerging Topics in Computing
- Vol. 13 (3) , 714-723
Fault Tolerance in Triplet Network Training: Analysis, Evaluation and Protection Methods
This paper investigates the tolerance of Triplet Networks (TNs) with a focus on faults in the training process. For compatibility with the existing literature. So-called stuck-at faults of a functional nature are considered for the operation of the neurons and activation function. While TNs are shown to be generally robust against such faults in the anchor and positive subnetworks, the presented analysis reveals a significant vulnerability in the negative subnetwork, in which stuck-at faults can lead to false convergence and training failures. An in-depth treatment is provided to show the incorrect convergence of training in the presence of stuck-at faults, highlighting the behavior of the network with faulty neurons. Extensive simulations are presented to evaluate the impact of these faults and propose two innovative fault-tolerant methods: the regularization of the anchor outputs and the modified margin. Simulation shows that false convergence can be very efficiently avoided by utilizing the proposed techniques, and thus the overall accuracy loss of the TN is negligible. These findings contribute to the understanding of fault tolerance in emerging neural networks such as TNs and offer practical solutions for enhancing their robustness against faults.
- 19 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Circuits and Systems Magazine
- Vol. 24 (3) , 19-33
Chiplet-GAN: Chiplet-Based Accelerator Design for Scalable Generative Adversarial Network Inference [Feature]
Generative adversarial networks (GANs) have emerged as a powerful solution for generating synthetic data when the availability of large, labeled training datasets is limited or costly in large-scale machine learning systems. Recent advancements in GAN models have extended their applications across diverse domains, including medicine, robotics, and content synthesis. These advanced GAN models have gained recognition for their excellent accuracy by scaling the model. However, existing accelerators face scalability challenges when dealing with large-scale GAN models. As the size of GAN models increases, the demand for computation and communication resources during inference continues to grow. To address this scalability issue, this article proposes Chiplet-GAN, a chiplet-based accelerator design for GAN inference. Chiplet-GAN enables scalability by adding more chiplets to the system, thereby supporting the scaling of computation capabilities. To handle the increasing communication demand as the system and model scale, a novel interconnection network with adaptive topology and passive/active network links is developed to provide adequate communication support for Chiplet-GAN. Coupled with workload partition and allocation algorithms, Chiplet-GAN reduces execution time and energy consumption for GAN inference workloads as both model and chiplet-system scales. Evaluation results using various GAN models show the effectiveness of Chiplet-GAN. On average, compared to GANAX, SpAtten, and Simba, the Chiplet-GAN reduces execution time and energy consumption by 34% and 21%, respectively. Furthermore, as the system scales for large-scale GAN model inference, Chiplet-GAN achieves reductions in execution time of up to 63% compared to the Simba, a chiplet-based accelerator.
- 19 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Circuits and Systems Magazine
- Vol. 24 (3) , 19-33
Chiplet-GAN: Chiplet-Based Accelerator Design for Scalable Generative Adversarial Network Inference [Feature]
Generative adversarial networks (GANs) have emerged as a powerful solution for generating synthetic data when the availability of large, labeled training datasets is limited or costly in large-scale machine learning systems. Recent advancements in GAN models have extended their applications across diverse domains, including medicine, robotics, and content synthesis. These advanced GAN models have gained recognition for their excellent accuracy by scaling the model. However, existing accelerators face scalability challenges when dealing with large-scale GAN models. As the size of GAN models increases, the demand for computation and communication resources during inference continues to grow. To address this scalability issue, this article proposes Chiplet-GAN, a chiplet-based accelerator design for GAN inference. Chiplet-GAN enables scalability by adding more chiplets to the system, thereby supporting the scaling of computation capabilities. To handle the increasing communication demand as the system and model scale, a novel interconnection network with adaptive topology and passive/active network links is developed to provide adequate communication support for Chiplet-GAN. Coupled with workload partition and allocation algorithms, Chiplet-GAN reduces execution time and energy consumption for GAN inference workloads as both model and chiplet-system scales. Evaluation results using various GAN models show the effectiveness of Chiplet-GAN. On average, compared to GANAX, SpAtten, and Simba, the Chiplet-GAN reduces execution time and energy consumption by 34% and 21%, respectively. Furthermore, as the system scales for large-scale GAN model inference, Chiplet-GAN achieves reductions in execution time of up to 63% compared to the Simba, a chiplet-based accelerator.
- 26 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4638-4651
A Balanced Sparse Matrix Convolution Accelerator for Efficient CNN Training
Sparse Convolutional Neural Network (CNN) training is well known to be time-consuming due to significant off-chip memory traffic. To effectively deploy sparse training, existing accelerators store matrices in a compressed format to eliminate memory accesses for zeros; hence, accelerators are designed to process compressed matrices to avoid zero computations. We have observed that the compression rate is greatly affected by the sparsity in the matrices with different formats. Given the varying levels of sparsity in activations, weights, errors, and gradients matrices throughout the sparse training process, it becomes impractical to achieve consistently high compression rates using a singular compression method for the entire duration of the training. Moreover, random zeros in the matrices result in irregular computation patterns, further increasing execution time. To address these issues, we propose a balanced sparse matrix convolution accelerator design for efficient CNN training. Specifically, a dual matrix compression technique is developed that seamlessly combines two widely used sparse matrix compression formats with a control algorithm for lower memory traffic during training. Based on this compression technique, a two-level workload balancing technique is then designed to further reduce the execution time and energy consumption. Finally, an accelerator is implemented to support the proposed techniques. The cycle-accurate simulation results show that the proposed accelerator reduces the execution time by 34% and the energy consumption by 24% on average compared to existing sparse training accelerators.
- 26 July 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems I: Regular Papers
- Vol. 71 (10) , 4638-4651
A Balanced Sparse Matrix Convolution Accelerator for Efficient CNN Training
Sparse Convolutional Neural Network (CNN) training is well known to be time-consuming due to significant off-chip memory traffic. To effectively deploy sparse training, existing accelerators store matrices in a compressed format to eliminate memory accesses for zeros; hence, accelerators are designed to process compressed matrices to avoid zero computations. We have observed that the compression rate is greatly affected by the sparsity in the matrices with different formats. Given the varying levels of sparsity in activations, weights, errors, and gradients matrices throughout the sparse training process, it becomes impractical to achieve consistently high compression rates using a singular compression method for the entire duration of the training. Moreover, random zeros in the matrices result in irregular computation patterns, further increasing execution time. To address these issues, we propose a balanced sparse matrix convolution accelerator design for efficient CNN training. Specifically, a dual matrix compression technique is developed that seamlessly combines two widely used sparse matrix compression formats with a control algorithm for lower memory traffic during training. Based on this compression technique, a two-level workload balancing technique is then designed to further reduce the execution time and energy consumption. Finally, an accelerator is implemented to support the proposed techniques. The cycle-accurate simulation results show that the proposed accelerator reduces the execution time by 34% and the energy consumption by 24% on average compared to existing sparse training accelerators.