Aug 2026· International Journal of Research & Technology· 0 citations· 2 references
TL;DR
Findings confirm that combining complementary compression strategies yields substantially better performance-efficiency trade-offs than any single technique applied in isolation.
Abstract
The rapid growth of deep neural networks has produced models that deliver state-of-the-art accuracy while imposing substantial computational, memory, and energy costs, which limits their deployment on resource-constrained platforms such as mobile devices, embedded systems, and edge accelerators. This paper investigates model compression and optimization strategies, namely magnitude-based and structured pruning, post-training and quantization-aware training, and response-based knowledge distillation, as mechanisms for improving the performance-efficiency trade-off of convolutional neural networks. A unified experimental framework is proposed in which a ResNet-based baseline model is compressed using each technique independently and in hybrid combination, and the resulting models are evaluated on accuracy, model size, floating-point operations (FLOPs), inference latency, and energy consumption. Experimental results, summarized across six comparative tables and four analytical figures, indicate that structured pruning at moderate sparsity levels (30-50%) preserves accuracy within 1-2 percentage points of the baseline while reducing model size by up to 2.6x, that 8-bit quantization achieves near-lossless compression with a 4x reduction in memory footprint, and that a hybrid pipeline combining pruning, quantization, and distillation achieves an 8.5x compression ratio with an accuracy degradation of under 5%. These findings confirm that combining complementary compression strategies yields substantially better performance-efficiency trade-offs than any single technique applied in isolation. The paper concludes with a discussion of open challenges, including hardware-aware compression, automated compression policy search, and compression of large-scale transformer architectures, and outlines directions for future research.
The proposed hybrid pipeline includes structured pruning, INT8 quantization and task-specific knowledge distillation, which is benchmarked against standalone methods and reinforces the idea of upper bound projection based approach for accuracy-oriented, multi-level compression.
Krishna Kumar Tiwari, Komal Tahiliani, Uma Shankar Birthare et al.· International journal of com...· 0 citations
Experimental results demonstrate that effective compression significantly reduces model size and computational cost with minimal performance loss, highlighting the importance of compression-aware design and concluding as a valuable reference for building efficient and scalable AI systems.
Daniel Rodríguez· International Journal of App...· 0 citations
This paper proposes a novel compression method that uses Singular Value Decomposition (SVD) informed by data examples that achieves substantial reductions in model size and computation with minimal impact on accuracy and demonstrates its applicability in compressing deep neural networks to a fraction of their original size.
Abdelfattah Toulaoui, H. Khalfi, I. Hafidi· Journal of Artificial Intell...· 0 citations
The rapid growth of deep learning models has increased computational requirements, energy consumption, and carbon emissions, raising concerns about the sustainability of artificial intelligence systems. Existing compression techniques, such as pruning and quantization, are typically evaluated in isolation and provide limited guidance for selecting energy-efficient configurations. To address this challenge, this paper presents a Green AI framework that predicts and recommends model compression strategies based on their expected efficiency gains within an empirically validated candidate space. The framework profiles pruning and quantization across nine architectures and employs supervised learning models to estimate their effects on memory usage, inference latency, energy consumption, and CO2 emissions. A recommendation module then ranks candidate configurations according to a user-specified optimization objective. The framework was evaluated using 276 profiling experiments spanning seven convolutional neural networks (ResNet18, ResNet50, VGG11, VGG16, MobileNet-V2, MobileNet-V3-Small, and EfficientNet-B0) and two Transformer models (BERT and DistilBERT) in a CPU-only deployment environment. Memory and latency prediction achieved R2 values above 0.95, while energy and CO2 reduction prediction achieved mean absolute errors of 3.27 and 2.85 percentage points, respectively, with R2 values of 0.71 and 0.73. Quantization generally produced larger energy and CO2 reductions than pruning, whereas pruning provided more moderate efficiency gains. However, six of the nine architectures exhibited increased runtime memory usage after compression, demonstrating that memory must be treated as an independent deployment constraint. Task-level evaluation further showed that INT8 quantization reduced BERT classification accuracy by only 1.3 percentage points. Results indicate that the optimal compression strategy is strongly architecture-dependent, highlighting the value of data-driven recommendation over fixed heuristic selection. Validation is performed against the closest experimentally observed configurations; extending the framework to previously unseen architectures remains an important direction for future work.
Adhithi C. Iyer, N. Akshatha, B. Suma· IEEE Access· 0 citations
The growing incorporation of deep learning into autonomous robotic systems has highlighted the importance of efficient inference on resource-constrained edge platforms. Moreover, License Plate Detection (LPD) is a key component of intelligent transportation systems, requiring high accuracy and real-time performance for practical deployment. However, deep learning detectors often remain computationally demanding for edge devices with limited resources. This paper proposes an edge-oriented optimization framework for YOLOv8-based LPD, combining structured pruning, INT8 quantization. Extensive experiments evaluate the trade-off between accuracy, computational complexity, and inference latency. Results show that combining pruning and quantization significantly reduce model size and computational cost while preserving detection performance. The optimized model achieves real-time inference exceeding 100 FPS on an Intel CPU, demonstrating its suitability for resource-constrained edge environments. These findings confirm that hybrid compression strategies enable efficient deployment of deep learning–based LPD systems in real-world applications.
S. Ameur, Amir Ismail, Xavier Clady et al.· International Conference on...· 0 citations
Extreme compression of deep neural networks, up to full binarization, dramatically reduces memory footprint and arithmetic complexity, facilitating deployment on constrained edge hardware with field-programmable gate arrays (FPGAs) and microcontrollers. Although combining binarization with pruning promises additional efficiency gains, existing pruning strategies are ill-suited to binarized representations and rarely translate into meaningful hardware savings. We introduce a PyTorch-based, research-oriented framework that incorporates freezing and pruning mechanisms for designing and optimizing binarized neural networks. The framework enables rapid and reproducible evaluation of state-of-the-art approaches and the fast prototyping of new ones. Leveraging this framework, we propose a novel pruning method that accounts for the relative importance of learned parameters across abstraction levels. Such a global weighting mechanism consistently achieves a superior trade-off between model accuracy and pruning rate, achieving a 70% pruning rate on VGG11 with constant accuracy, while state-of-the-art results reach only 41% in the binarized setting.