Deploying AI on Edge: Advancement and Challenges in Edge Intelligence by Tianyu Wang, Jinyang Guo, Bowen Zhang, Ge Yang, Dong Li

“…Although edge intelligence methods have been proposed to alleviate the computational and storage burdens, they still face multiple persistent challenges, such as large-scale model deployment, poor interpretability, privacy and security vulnerabilities, and energy efficiency constraints. This article systematically reviews the current advancements in edge intelligence technologies, highlights key enabling techniques including model sparsity, quantization, knowledge distillation, neural architecture search, and federated learning, and explores their applications in industrial, automotive, healthcare, and consumer domains. …”
Get full text

Benchmarking In-Sensor Machine Learning Computing: An Extension to the MLCommons-Tiny Suite by Fabrizio Maria Aymone, Danilo Pietro Pau

“…With the exponential growth of edge devices, efficient local processing is essential to mitigate economic costs, latency, and privacy concerns associated with the centralized cloud processing. …”
Get full text

An autonomous vehicles’ test case extraction method: Example of vehicle-to-pedestrian scenarios by Zhengping Tan, Qian Wang, Wenhao Hu, Pingfei Li, Liangliang Shi, Hao Feng

“…This paper proposes a novel method to enhance scenario adaptation by integrating quantization weights with a new clustering algorithm. …”
Get full text

Empowering Healthcare: TinyML for Precise Lung Disease Classification by Youssef Abadade, Nabil Benamar, Miloud Bagaa, Habiba Chaoui

“…Our data preprocessing included bandpass filtering and feature extraction through Mel-Frequency Cepstral Coefficients (MFCCs). We applied quantization techniques to ensure model efficiency. The custom CNN model achieved the highest performance, with 96% accuracy and 97% precision, recall, and F1-scores, while maintaining moderate resource usage. …”
Get full text

Lightweight deep learning method for end-to-end point cloud registration by Linjun Jiang, Yue Liu, Zhiyuan Dong, Yinghao Li, Yusong Lin

“…Specifically, our approach utilizes pruning and weight-sharing quantization techniques to reduce model size and simplify the network structure. …”
Get full text

Binary-Weighted Neural Networks Using FeRAM Array for Low-Power AI Computing by Seung-Myeong Cho, Jaesung Lee, Hyejin Jo, Dai Yun, Jihwan Moon, Kyeong-Sik Min

“…Furthermore, the combination of binary weight quantization and in-memory computing enables energy-efficient inference without significant loss in recognition accuracy, as validated using MNIST datasets. …”
Get full text

A DSP–FPGA Heterogeneous Accelerator for On-Board Pose Estimation of Non-Cooperative Targets by Qiuyu Song, Kai Liu, Shangrong Li, Mengyuan Wang, Junyi Wang

“…Optimization techniques, including batch normalization fusion, ReLU integration, and linear quantization, are applied to improve inference efficiency. …”
Get full text

CNN-Based Optimization for Fish Species Classification: Tackling Environmental Variability, Class Imbalance, and Real-Time Constraints by Amirhosein Mohammadisabet, Raza Hasan, Vishal Dattana, Salman Mahmood, Saqib Hussain

“…Optimization techniques, including pruning and quantization, reduced model size by 73.7%, enabling real-time deployment on resource-constrained devices. …”
Get full text

A seed-epidermis-feature-recognition-based lightweight peanut seed selection method for embedded systems by Dehao Li, Jinlong Huang, Xincheng Li, Zhaolei Yang, Xueke An, Pengfei Xu, Yuliang Yun

“…The model was based on deep learning and model quantization techniques. The transfer learning method was used to use four pre-trained models, EfficientNet_b0, EfficientNetv2-b0, MobileNet_v2_35_224, and NasNet_Mobile, as feature extraction layers, the input layer was added before the feature extraction layer, and the dropout and dense layers were added after the feature extraction layer to construct a classifier. …”
Get full text

Trends and technologies of video coding by Jiantong ZHOU, Haitao YANG, Dong LIU, Xiang MA, Tian WANG

“…The current video coding uses block based hybrid architecture,which uses predictive,transform,quantization and entropy coding techniques to efficiently compress video signals.Further optimizations on current architectures provide more flexible processing and coding for local characteristics of video image signals.Video coding based on machine learning was expected to change the existing hybrid coding framework partially or comprehensively,and bring new research ideas to video coding.In addition to existing 2D video signal,the future of video also needs to spherical video coding and volumetric video coding for AR/VR applications,the new video source data format of the video encoding technology has brought new opportunities and challenges.…”
Get full text

A Comparative Study and Optimization of Camera-Based BEV Segmentation for Real-Time Autonomous Driving by Woomin Jun, Sungjin Lee

“…Notably, the lift–splat–shoot view transformation model with the InternImage-T encoder and EfficientNet-B0 decoder demonstrated performance of 53.1 mIoU while achieving high efficiency (51.7 ms and 159.5 MB, respectively). …”
Get full text

VVC coded distortion prediction model based on frame-level transform coefficient modeling of generalized Gaussian distribution by Yiyin GU, Hongkui WANG, Haibin YIN

“…In versatile video coding (VVC), a variety of advanced coding tools work together to achieve excellent coding performance.Compared with high efficient video coding (HEVC), the transform coefficient distribution (TCD) of VVC has sharper peaks.In order to solve this phenomenon, the probability density function (PDF) of frame-level TCD was modeled, and a frame-level coding distortion prediction model based on statistical modeling was proposed, which modeled frame-level distortion as a function of TCD distribution parameters and quantization parameters.The experimental results show that compared with the Laplace distribution and Cauchy distribution, the generalized Gaussian distribution has the best performance in TCD probability density fitting.The prediction results based on the generalized Gaussian distribution distortion prediction model are closest to the actual coding distortion.…”
Get full text

Ensemble Transformer–Based Detection of Fake and AI–Generated News by Md. Ishraquzzaman, Mohammed Ashraful Islam Chowdhury, Shahreen Rahman, Riasat Khan

“…The proposed ensemble model is optimized by applying model pruning (reducing parameters from 265M to 210M, improving training time by 25%) and dynamic quantization (reducing model size by 50%, maintaining 95.68% accuracy), enhancing scalability and efficiency while minimizing computational overhead. …”
Get full text

Automated Arrhythmia Classification System: Proof-of-Concept With Lightweight Model on an Ultra-Edge Device by Namho Kim, Seongjae Lee, Seungmin Kim, Sung-Min Park

“…Model compression methods including knowledge distillation, pruning, and quantization were employed to enhance arrhythmia classification performance while reducing computational complexity. …”
Get full text

Lightweight Deep Learning Model for Fire Classification in Tunnels by Shakhnoza Muksimova, Sabina Umirzakova, Jushkin Baltayev, Young-Im Cho

“…Deployment optimizations, such as quantization and layer fusion, ensure computational efficiency, achieving an average inference time of 12ms/frame, making it suitable for resource-constrained environments like IoT and edge devices. …”
Get full text

Analysis of space labeling through binary fingerprinting by Marouan Mizmizi, Luca Reggiani

“…In the context of fingerprinting applications, this article presents the performance analysis of a type of space labeling based on the binary quantization of the received signal strength indicator. …”
Get full text

Slim-sugarcane: a lightweight and high-precision method for sugarcane node detection and edge deployment in natural environments by Lijiao Wei, Lijiao Wei, Shuo Wang, Xinwei Liang, Dongjie Du, Xinyi Huang, Ming Li, Yuangang Hua, Weihua Huang, Zhenhui Zheng, Zhenhui Zheng

“…The proposed framework is optimized with TensorRT and deployed using FP16 quantization on the NVIDIA Jetson Orin NX platform to ensure real-time performance under limited hardware conditions. …”
Get full text

Category semantic and global relation distillation for object detection by Yanpeng LIANG, Zhonggui MA, Zongjie WANG, Zhuo LI

“…Compared with other baseline methods, the proposed approach achieved competitive improvements in mean average precision without considerably increasing the number of parameters and FLOPS during distillation training, thereby striking a better balance between accuracy and efficiency.…”
Get full text

ConflLlama: Domain-specific adaptation of large language models for conflict event classification by Shreyas Meher, Patrick T. Brandt

“…We demonstrate how to adapt open-source language models to specialized political science tasks, using conflict event classification as our proof of concept. Through quantization and efficient fine-tuning techniques, we show state-of-the-art performance while minimizing computational requirements. …”
Get full text

A Deep Learning-Driven Solution to Limited-Feedback MIMO Relaying Systems by Kwadwo Boateng Ofori-Amanfo, Bridget Durowaa Antwi-Boasiako, Prince Anokye, Suho Shin, Kyoung-Jae Lee

“…By harnessing binary feedback information from the implemented DNNs together with efficient beamforming vectors, a novel approach to the resulting problem is presented. …”
Get full text