Ahmed M. Eltawil

Wenzhe Guo, Mohammed E Fouda, Ahmed M. Eltawil et al.

Empowered by the backpropagation (BP) algorithm, deep neural networks have dominated the race in solving various cognitive tasks. The restricted training pattern in the standard BP requires end-to-end error propagation, causing large memory cost and prohibiting model parallelization. Existing local training methods aim to resolve the training obstacle by completely cutting off the backward path between modules and isolating their gradients to reduce memory cost and accelerate the training process. These methods prevent errors from flowing between modules and hence information exchange, resulting in inferior performance. This work proposes a novel local training algorithm, BackLink, which introduces inter-module backward dependency and allows errors to flow between modules. The algorithm facilitates information to flow backward along with the network. To preserve the computational advantage of local training, BackLink restricts the error propagation length within the module. Extensive experiments performed in various deep convolutional neural networks demonstrate that our method consistently improves the classification performance of local training algorithms over other methods. For example, in ResNet32 with 16 local modules, our method surpasses the conventional greedy local training method by 4.00\% and a recent work by 1.83\% in accuracy on CIFAR10, respectively. Analysis of computational costs reveals that small overheads are incurred in GPU memory costs and runtime on multiple GPUs. Our method can lead up to a 79\% reduction in memory cost and 52\% in simulation runtime in ResNet110 compared to the standard BP. Therefore, our method could create new opportunities for improving training algorithms towards better efficiency and biological plausibility.

Diego A. Silva, Kamilya Smagulova, Ahmed Elsheikh et al.

Object detection is crucial in various cutting-edge applications, such as autonomous vehicles and advanced robotics systems, primarily relying on data from conventional frame-based RGB sensors. However, these sensors often struggle with issues like motion blur and poor performance in challenging lighting conditions. In response to these challenges, event-based cameras have emerged as an innovative paradigm. These cameras, mimicking the human eye, demonstrate superior performance in environments with fast motion and extreme lighting conditions while consuming less power. This study introduces ReYOLOv8, an advanced object detection framework that enhances a leading frame-based detection system with spatiotemporal modeling capabilities. We implemented a low-latency, memory-efficient method for encoding event data to boost the system's performance. We also developed a novel data augmentation technique tailored to leverage the unique attributes of event data, thus improving detection accuracy. Our models outperformed all comparable approaches in the GEN1 dataset, focusing on automotive applications, achieving mean Average Precision (mAP) improvements of 5%, 2.8%, and 2.5% across nano, small, and medium scales, respectively.These enhancements were achieved while reducing the number of trainable parameters by an average of 4.43% and maintaining real-time processing speeds between 9.2ms and 15.5ms. On the PEDRo dataset, which targets robotics applications, our models showed mAP improvements ranging from 9% to 18%, with 14.5x and 3.8x smaller models and an average speed enhancement of 1.67x.

Abdulkadir Celik, Ahmed M. Eltawil

The majority of data-driven wireless research leans heavily on discriminative AI (DAI) that requires vast real-world datasets. Unlike the DAI, Generative AI (GenAI) pertains to generative models (GMs) capable of discerning the underlying data distribution, patterns, and features of the input data. This makes GenAI a crucial asset in wireless domain wherein real-world data is often scarce, incomplete, costly to acquire, and hard to model or comprehend. With these appealing attributes, GenAI can replace or supplement DAI methods in various capacities. Accordingly, this combined tutorial-survey paper commences with preliminaries of 6G and wireless intelligence by outlining candidate 6G applications and services, presenting a taxonomy of state-of-the-art DAI models, exemplifying prominent DAI use cases, and elucidating the multifaceted ways through which GenAI enhances DAI. Subsequently, we present a tutorial on GMs by spotlighting seminal examples such as generative adversarial networks, variational autoencoders, flow-based GMs, diffusion-based GMs, generative transformers, large language models, to name a few. Contrary to the prevailing belief that GenAI is a nascent trend, our exhaustive review of approximately 120 technical papers demonstrates the scope of research across core wireless research areas, including physical layer design; network optimization, organization, and management; network traffic analytics; cross-layer network security; and localization & positioning. Furthermore, we outline the central role of GMs in pioneering areas of 6G network research, including semantic/THz/near-field communications, ISAC, extremely large antenna arrays, digital twins, AI-generated content services, mobile edge computing and edge AI, adversarial ML, and trustworthy AI. Lastly, we shed light on the multifarious challenges ahead, suggesting potential strategies and promising remedies.

Ahmad M. Nazar, Abdulkadir Celik, Asmaa Abdallah et al.

Maintaining robust millimeter-wave (mmWave) connectivity in vehicular networks requires real-time adaptation to environmental dynamics, sensor degradation, and link variability. This paper presents Enwar 3.0, an environment-aware reasoning framework that unifies multi-modal sensing, agentic large language models (LLMs), and context-driven model selection for predictive beamforming, blockage detection, and handover management. Building upon prior iterations of Enwar, the proposed architecture integrates a classifier-driven assessment of sensor health with a primed LLM that orchestrates multiple specialized agents through structured, task-aware prompting. A novel synthetic degradation pipeline enables the training of a sensor degradation classifier that detects real-time impairments across camera, radar, LiDAR, and GPS inputs, achieving over 99% accuracy. The LLM, trained via chain-of-thought (CoT) priming and human-in-the-loop feedback, coordinates agent calls for beam selection, blockage forecasting, and environment perception while dynamically loading sensor-specific models based on environmental context. Extensive evaluations across 15 sensor combinations demonstrate that Enwar 3.0 delivers state-of-the-art performance in both predictive accuracy and interpretability, with beam selection accuracy exceeding 88%, blockage F1-scores surpassing 98%, and reasoning correctness reaching 87% on complex decision prompts. This work establishes a scalable foundation for LLM-integrated wireless systems that reason, perceive, and adapt in real-time.

Nasir Khan, Asmaa Abdallah, Abdulkadir Celik et al.

Artificial intelligence (AI) is expected to significantly enhance radio resource management (RRM) in sixth-generation (6G) networks. However, the lack of explainability in complex deep learning (DL) models poses a challenge for practical implementation. This paper proposes a novel explainable AI (XAI)- based framework for feature selection and model complexity reduction in a model-agnostic manner. Applied to a multi-agent deep reinforcement learning (MADRL) setting, our approach addresses the joint sub-band assignment and power allocation problem in cellular vehicle-to-everything (V2X) communications. We propose a novel two-stage systematic explainability framework leveraging feature relevance-oriented XAI to simplify the DRL agents. While the former stage generates a state feature importance ranking of the trained models using Shapley additive explanations (SHAP)-based importance scores, the latter stage exploits these importance-based rankings to simplify the state space of the agents by removing the least important features from the model input. Simulation results demonstrate that the XAI-assisted methodology achieves 97% of the original MADRL sum-rate performance while reducing optimal state features by 28%, average training time by 11%, and trainable weight parameters by 46% in a network with eight vehicular pairs.

Mariam Rakka, Rachid Karami, Ahmed M. Eltawil et al.

Mixed-precision quantization works Neural Networks (NNs) are gaining traction for their efficient realization on the hardware leading to higher throughput and lower energy. In-Memory Computing (IMC) accelerator architectures are offered as alternatives to traditional architectures relying on a data-centric computational paradigm, diminishing the memory wall problem, and scoring high throughput and energy efficiency. These accelerators can support static fixed-precision but are not flexible to support mixed-precision NNs. In this paper, we present BF-IMNA, a bit fluid IMC accelerator for end-to-end Convolutional NN (CNN) inference that is capable of static and dynamic mixed-precision without any hardware reconfiguration overhead at run-time. At the heart of BF-IMNA are Associative Processors (APs), which are bit-serial word-parallel Single Instruction, Multiple Data (SIMD)-like engines. We report the performance of end-to-end inference of ImageNet on AlexNet, VGG16, and ResNet50 on BF-IMNA for different technologies (eNVM and NVM), mixed-precision configurations, and supply voltages. To demonstrate bit fluidity, we implement HAWQ-V3's per-layer mixed-precision configurations for ResNet18 on BF-IMNA using different latency budgets, and results reveal a trade-off between accuracy and Energy-Delay Product (EDP): On one hand, mixed-precision with a high latency constraint achieves the closest accuracy to fixed-precision INT8 and reports a high (worse) EDP compared to fixed-precision INT4. On the other hand, with a low latency constraint, BF-IMNA reports the closest EDP to fixed-precision INT4, with a higher degradation in accuracy compared to fixed-precision INT8. We also show that BF-IMNA with fixed-precision configuration still delivers performance that is comparable to current state-of-the-art accelerators: BF-IMNA achieves $20\%$ higher energy efficiency and $2\%$ higher throughput.

Nasir Khan, Asmaa Abdallah, Abdulkadir Celik et al.

In line with the AI-native 6G vision, explainability and robustness are crucial for building trust and ensuring reliable performance in millimeter-wave (mmWave) systems. Efficient beam alignment is essential for initial access, but deep learning (DL) solutions face challenges, including high data collection overhead, hardware constraints, lack of explainability, and susceptibility to adversarial attacks. This paper proposes a robust and explainable DL-based beam alignment engine (BAE) for mmWave multiple-input multiple output (MIMO) systems. The BAE uses received signal strength indicator (RSSI) measurements from wide beams to predict the best narrow beam, reducing the overhead of exhaustive beam sweeping. To overcome the challenge of real-world data collection, this work leverages a site-specific digital twin (DT) to generate synthetic channel data closely resembling real-world environments. A model refinement via transfer learning is proposed to fine-tune the pre-trained model residing in the DT with minimal real-world data, effectively bridging mismatches between the digital replica and real-world environments. To reduce beam training overhead and enhance transparency, the framework uses deep Shapley additive explanations (SHAP) to rank input features by importance, prioritizing key spatial directions and minimizing beam sweeping. It also incorporates the Deep k-nearest neighbors (DkNN) algorithm, providing a credibility metric for detecting out-of-distribution inputs and ensuring robust, transparent decision-making. Experimental results show that the proposed framework reduces real-world data needs by 70%, beam training overhead by 62%, and improves outlier detection robustness by up to 8.5x, achieving near-optimal spectral efficiency and transparent decision making compared to traditional softmax based DL models.

Nasir Khan, Asmaa Abdallah, Abdulkadir Celik et al.

Integrated artificial intelligence (AI) and communication has been recognized as a key pillar of 6G and beyond networks. In line with AI-native 6G vision, explainability and robustness in AI-driven systems are critical for establishing trust and ensuring reliable performance in diverse and evolving environments. This paper addresses these challenges by developing a robust and explainable deep learning (DL)-based beam alignment engine (BAE) for millimeter-wave (mmWave) multiple-input multiple-output (MIMO) systems. The proposed convolutional neural network (CNN)-based BAE utilizes received signal strength indicator (RSSI) measurements over a set of wide beams to accurately predict the best narrow beam for each UE, significantly reducing the overhead associated with exhaustive codebook-based narrow beam sweeping for initial access (IA) and data transmission. To ensure transparency and resilience, the Deep k-Nearest Neighbors (DkNN) algorithm is employed to assess the internal representations of the network via nearest neighbor approach, providing human-interpretable explanations and confidence metrics for detecting out-of-distribution inputs. Experimental results demonstrate that the proposed DL-based BAE exhibits robustness to measurement noise, reduces beam training overhead by 75% compared to the exhaustive search while maintaining near-optimal performance in terms of spectral efficiency. Moreover, the proposed framework improves outlier detection robustness by up to 5x and offers clearer insights into beam prediction decisions compared to traditional softmax-based classifiers.

Diego A. Silva, Ahmed Elsheikh, Kamilya Smagulova et al.

Event-based cameras are sensors that simulate the human eye, offering advantages such as high-speed robustness and low power consumption. Established Deep Learning techniques have shown effectiveness in processing event data. Chimera is a Block-Based Neural Architecture Search (NAS) framework specifically designed for Event-Based Object Detection, aiming to create a systematic approach for adapting RGB-domain processing methods to the event domain. The Chimera design space is constructed from various macroblocks, including Attention blocks, Convolutions, State Space Models, and MLP-mixer-based architectures, which provide a valuable trade-off between local and global processing capabilities, as well as varying levels of complexity. The results on the PErson Detection in Robotics (PEDRo) dataset demonstrated performance levels comparable to leading state-of-the-art models, alongside an average parameter reduction of 1.6 times.

Ahmet M. Elbir, Abdulkadir Celik, Asmaa Abdallah et al.

Next generation wireless networks focus on improving spectral efficiency (SE) while reducing power consumption and hardware cost. Reconfigurable intelligent surfaces (RISs) offer a viable solution to meet these requirements. In order to enhance the SE, index modulation (IM) has been regarded as one of the enabling technologies via the transmission of additional information bits over the transmission media such as subcarriers, antennas and spatial paths. In this work, we explore the usage of spatial paths and introduce spatial path IM (SPIM) for RIS-aided massive multiple-input multiple-output (mMIMO) systems. Thus, the proposed framework improves the network efficiency and the coverage with the use of RIS while SPIM provides SE improvement. In order to perform SPIM, we exploit the spatial diversity of the millimeter wave channel and assign the index bits to the spatial patterns of the channel between the base station and the users through RIS. We introduce a low complexity approach for the design of hybrid beamformers, which are constructed by the steering vectors corresponding to the selected spatial path indices for SPIM-mMIMO. Furthermore, we conduct a theoretical analysis on the SE of the proposed SPIM approach, and derive the SE relationship between the SPIM-based hybrid beamforming and fully digital (FD) beamforming. Via numerical simulations, we validate our theoretical results and show that the proposed SPIM approach presents an improved SE performance, even higher than that of the use of FD beamformers while using a few RF chains.

Mariam Rakka, Marios Fournarakis, Olga Krestinskaya et al.

The rapid scaling of language models (LMs) has resulted in unprecedented computational, memory, and energy requirements, making their training and deployment increasingly unsustainable. Quantization has emerged as an essential compression technique to reduce model size, alleviate memory bottlenecks, and accelerate inference. However, while uniform low-bit quantization (e.g., INT8, INT4) provides significant efficiency gains, it can degrade accuracy in sensitive components of transformer-based LMs. Mixed-precision quantization offers a promising alternative by selectively allocating precision across layers or within tensors to balance efficiency and accuracy. This survey provides a comprehensive overview of Mixed-Precision quantization frameworks for LMs (MXPLMs). We first review quantization fundamentals, including uniform and non-uniform quantizers, quantization granularity, and methods widely used in post-training quantization. We then categorize and compare recent MXPLM frameworks according to their bit allocation strategies and precision configurations across weights, activations, and key-value caches. A comparative analysis highlights differences in perplexity, zero-shot task performance, and deployment trade-offs. Furthermore, we contrast MXPLMs with earlier mixed-precision quantization methods for deep neural networks, identifying strategies that transfer and those that face challenges in the LM setting. Finally, we summarize open issues and future directions, including hardware-aware design, activation quantization, and scalable optimization methods for billion-parameter models. By consolidating recent advances, this work serves as a reference for understanding the current landscape and research prospects of mixed-precision quantization for large-scale language models.

Nasir Khan, Asmaa Abdallah, Abdulkadir Celik et al.

Efficient and reliable beam alignment is a critical requirement for mmWave multiple-input multiple-output (MIMO) systems, especially in 6G and beyond, where communication must be fast, adaptive, and resilient to real-world uncertainties. Existing deep learning (DL)-based beam alignment methods often neglect the underlying causal relationships between inputs and outputs, leading to limited interpretability, poor generalization, and unnecessary beam sweeping overhead. In this work, we propose a causally-aware DL framework that integrates causal discovery into beam management pipeline. Particularly, we propose a novel two-stage causal beam selection algorithm to identify a minimal set of relevant inputs for beam prediction. First, causal discovery learns a Bayesian graph capturing dependencies between received power inputs and the optimal beam. Then, this graph guides causal feature selection for the DL-based classifier. Simulation results reveal that the proposed causal beam selection matches the performance of conventional methods while drastically reducing input selection time by 94.4% and beam sweeping overhead by 59.4% by focusing only on causally relevant features.

Ahmad M. Nazar, Abdulkadir Celik, Mohamed Y. Selim et al.

Vehicular communication systems operating in the millimeter wave (mmWave) band are highly susceptible to signal blockage from dynamic obstacles such as vehicles, pedestrians, and infrastructure. To address this challenge, we propose a proactive blockage prediction framework that utilizes multi-modal sensing, including camera, GPS, LiDAR, and radar inputs in an infrastructure-to-vehicle (I2V) setting. This approach uses modality-specific deep learning models to process each sensor stream independently and fuses their outputs using a softmax-weighted ensemble strategy based on validation performance. Our evaluations, for up to 1.5s in advance, show that the camera-only model achieves the best standalone trade-off with an F1-score of 97.1% and an inference time of 89.8ms. A camera+radar configuration further improves accuracy to 97.2% F1 at 95.7ms. Our results display the effectiveness and efficiency of multi-modal sensing for mmWave blockage prediction and provide a pathway for proactive wireless communication in dynamic environments.

Asmaa Abdallah, Abdullatif Albaseer, Abdulkadir Celik et al.

The transition to 6G networks promises unprecedented advancements in wireless communication, with increased data rates, ultra-low latency, and enhanced capacity. However, the complexity of managing and optimizing these next-generation networks presents significant challenges. The advent of large language models (LLMs) has revolutionized various domains by leveraging their sophisticated natural language understanding capabilities. However, the practical application of LLMs in wireless network orchestration and management remains largely unexplored. Existing literature predominantly offers visionary perspectives without concrete implementations, leaving a significant gap in the field. To address this gap, this paper presents NETORCHLLM, a wireless NETwork ORCHestrator LLM framework that uses LLMs to seamlessly orchestrate diverse wireless-specific models from wireless communication communities using their language understanding and generation capabilities. A comprehensive framework is introduced, demonstrating the practical viability of our approach and showcasing how LLMs can be effectively harnessed to optimize dense network operations, manage dynamic environments, and improve overall network performance. NETORCHLLM bridges the theoretical aspirations of prior research with practical, actionable solutions, paving the way for future advancements in integrating generative AI technologies within the wireless communications sector.

Wenzhe Guo, Mohammed E. Fouda, Ahmed M. Eltawil et al.

Directly training spiking neural networks (SNNs) has remained challenging due to complex neural dynamics and intrinsic non-differentiability in firing functions. The well-known backpropagation through time (BPTT) algorithm proposed to train SNNs suffers from large memory footprint and prohibits backward and update unlocking, making it impossible to exploit the potential of locally-supervised training methods. This work proposes an efficient and direct training algorithm for SNNs that integrates a locally-supervised training method with a temporally-truncated BPTT algorithm. The proposed algorithm explores both temporal and spatial locality in BPTT and contributes to significant reduction in computational cost including GPU memory utilization, main memory access and arithmetic operations. We thoroughly explore the design space concerning temporal truncation length and local training block size and benchmark their impact on classification accuracy of different networks running different types of tasks. The results reveal that temporal truncation has a negative effect on the accuracy of classifying frame-based datasets, but leads to improvement in accuracy on dynamic-vision-sensor (DVS) recorded datasets. In spite of resulting information loss, local training is capable of alleviating overfitting. The combined effect of temporal truncation and local training can lead to the slowdown of accuracy drop and even improvement in accuracy. In addition, training deep SNNs models such as AlexNet classifying CIFAR10-DVS dataset leads to 7.26% increase in accuracy, 89.94% reduction in GPU memory, 10.79% reduction in memory access, and 99.64% reduction in MAC operations compared to the standard end-to-end BPTT.

Melika Payvand, Mohammed E. Fouda, Fadi Kurdahi et al.

Recent breakthroughs in neuromorphic computing show that local forms of gradient descent learning are compatible with Spiking Neural Networks (SNNs) and synaptic plasticity. Although SNNs can be scalably implemented using neuromorphic VLSI, an architecture that can learn using gradient-descent in situ is still missing. In this paper, we propose a local, gradient-based, error-triggered learning algorithm with online ternary weight updates. The proposed algorithm enables online training of multi-layer SNNs with memristive neuromorphic hardware showing a small loss in the performance compared with the state of the art. We also propose a hardware architecture based on memristive crossbar arrays to perform the required vector-matrix multiplications. The necessary peripheral circuitry including pre-synaptic, post-synaptic and write circuits required for online training, have been designed in the sub-threshold regime for power saving with a standard 180 nm CMOS process.