Article
Computer Science, Information Systems
Israel Pineda, Dustin Carrion-Ojeda, Rigoberto Fonseca-Delgado
Summary: RADENN is a domain-specific language aimed at rapidly developing fully connected neural networks for classification and regression problems. It is built on top of Keras API with Tensorflow as its backend and incorporates specific data types and built-in functions to facilitate network creation, training, and evaluation. RADENN is an ideal tool for various professionals who need a fast and efficient way to create prototypes and models without extensive programming or deep learning knowledge. This work provides a detailed overview of RADENN's features and compares it with widely used libraries such as Keras and PyTorch.
Article
Computer Science, Information Systems
Enrico Russo, Maurizio Palesi, Salvatore Monteleone, Davide Patti, Andrea Mineo, Giuseppe Ascia, Vincenzo Catania
Summary: This article proposes a compression technique called LineCompress, which aims to reduce memory and communication overhead on resource-constrained IoT devices, thus improving the inference speed and energy efficiency of deep neural networks.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Artificial Intelligence
Qiang Gao, Zhipeng Luo, Diego Klabjan, Fengli Zhang
Summary: Continual learning with efficient architecture search (CLEAS) is proposed to overcome the challenges of catastrophic forgetting, adapting to new tasks, and controlling model complexity. By leveraging neuron-level NAS, CLEAS achieves higher classification accuracy by reusing old neurons and adding new ones on simpler neural architectures.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Information Systems
Marc Rothmann, Mario Porrmann
Summary: This paper presents a review of hardware architectures for accelerating reinforcement learning algorithms, focusing on FPGA-based implementations and considering GPU-based approaches as well. It compares the techniques employed in different implementations and suggests possible areas for future work.
Article
Optics
Jinhwa Gene, Jong Moo Sohn, Hyung Cheol Shin, Suntak Park
Summary: In this paper, we address the issue of optical diffractions caused by the binary nature of DMD operation in coherent optical information processing systems. We characterize the diffraction phenomena and propose a new DMD operation method and modified structure of the 4f-system. We demonstrate the performance of the optimized 4f-system by implementing high bit-depth image information processing.
Article
Engineering, Electrical & Electronic
Efe Camci, Manas Gupta, Min Wu, Jie Lin
Summary: In this paper, a novel method called QLP is proposed for pruning deep neural networks using deep Q-learning. The method achieves superior granular pruning by visiting each layer multiple times and pruning them little by little at each visit. It has the flexibility to execute a whole range of sparsity ratios for each layer, which enables aggressive pruning without compromising accuracy. Furthermore, the method features a simple, generic state definition and utilizes a carefully designed curriculum to deliver better accuracy at high sparsity levels.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Computer Science, Information Systems
Changchun Zhang, Junguo Zhang
Summary: This paper proposes a method called Transferable Regularization and Normalization (TRN) for unsupervised domain adaptation. TRN adjusts feature norms and improves normalization techniques to avoid negative transfer and facilitate positive transfer. Evaluation results show that TRN achieves state-of-the-art performance on multiple benchmark datasets.
INFORMATION SCIENCES
(2022)
Article
Multidisciplinary Sciences
Jieting Chen, Chao Qian, Jie Zhang, Yuetian Jia, Hongsheng Chen
Summary: The authors propose a generation-elimination framework that accurately forecasts inaccessible spectra by correlating spectra from different frequency bands without consulting structural information. This framework accelerates the unification of metasurface designs and enables versatile applications involving cross-wavelength information correlation. The study also introduces a dimensionality reduction approach to visualize the abstract correlated spectra data encoded in latent spaces.
NATURE COMMUNICATIONS
(2023)
Article
Engineering, Biomedical
Mohammadreza Naderi, Nader Karimi, Ali Emami, Shahram Shirani, Shadrokh Samavi
Summary: This study aims to improve the performance of conditional Generative Adversarial Networks (cGANs) in translating images by learning the target domain distribution from limited data with the help of noise input. The proposed method achieves better model generalization and comparable results compared to state-of-the-art methods.
BIOMEDICAL SIGNAL PROCESSING AND CONTROL
(2023)
Article
Telecommunications
Fang Jiang, Da-Wei Chang, Song Ma, Yan-Jun Hu, Yao-Hua Xu
Summary: Sparse code multiple access (SCMA) is a technology proposed for large-scale intelligent terminal devices with high spectrum utilization. In this study, we design a new end-to-end autoencoder combining convolutional neural networks (CNNs) and residual networks to improve the accuracy and computational complexity of SCMA for the internet of things (IoT) scenario. Our scheme, with a residual network utilizing multitask learning and CNN units for SCMA codeword mapping, outperforms existing autoencoder schemes in terms of bit error rate (BER) and computational complexity according to simulations.
IEEE COMMUNICATIONS LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
M. Khojaste-Sarakhsi, Seyedhamidreza Shahabi Haghighi, S. M. T. Fatemi Ghomi, Elena Marchiori
Summary: Alzheimer's Disease (AD) is an irreversible neurodegenerative disease that leads to a gradual decline in cognitive abilities. The early detection of AD is challenging due to subtle changes in biomarkers, which are primarily detectable in different neuroimaging techniques. This survey examines approximately 100 published papers since 2019 that utilize deep learning models such as CNN, RNN, and generative models for AD diagnosis. Additionally, it investigates around 60 papers that apply trending topics or architectures for AD research. The challenges in this field are categorized and explained in terms of data, methodology, and clinical adoption. The paper concludes by discussing future perspectives and providing recommendations for further studies in AD diagnosis.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2022)
Article
Computer Science, Artificial Intelligence
Yuqiao Liu, Yanan Sun, Bing Xue, Mengjie Zhang, Gary G. Yen, Kay Chen Tan
Summary: Deep neural networks have achieved great success in many applications, but their architectures require labor-intensive and expert-designed processes. Neural architecture search (NAS) technology enables automatic design of architectures, with evolutionary computation (EC) methods gaining attention and success. However, there is currently no comprehensive summary of EC-based NAS algorithms.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Computer Science, Information Systems
Anthony Berthelier, Thierry Chateau, Stefan Duffner, Christophe Garcia, Christophe Blanc
Summary: This paper surveys methods suitable for porting deep neural networks on resource-limited devices, especially for smart cameras, which can be roughly divided into compression techniques and architecture optimization. Compression techniques include knowledge distillation, pruning, quantization, hashing, reduction of numerical precision and binarization, while architecture optimization focuses on enhancing network structures and neural architecture search techniques.
JOURNAL OF SIGNAL PROCESSING SYSTEMS FOR SIGNAL IMAGE AND VIDEO TECHNOLOGY
(2021)
Article
Computer Science, Artificial Intelligence
Juan Carlos Cepeda-Pacheco, Mari Carmen Domingo
Summary: We propose a tourist attraction IoT-enabled deep learning-based recommendation system to enhance tourist experience in a smart city. The system takes into account personal input features and real-time context information to recommend suitable tourist activities and attractions, resulting in an improved tourist experience.
NEURAL COMPUTING & APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Unai Garciarena, Alexander Mendiburu, Roberto Santana
Summary: Multi-task learning involves using a single model to perform multiple similar tasks, expanding performance range. This study introduces heterogeneous tasks in a single learning procedure and develops an illustrative model with classification, regression, and data sampling tasks. The model exhibits good performance, with potential for future research directions.
ACM TRANSACTIONS ON KNOWLEDGE DISCOVERY FROM DATA
(2021)