ETRI Journal最新文献

As computing systems become increasingly larger, high-performance computing (HPC) is gaining importance. In particular, as hyperscale artificial intelligence (AI) applications, such as large language models emerge, HPC has become important even in the field of AI. Important operations in hyperscale AI and HPC are mainly linear algebraic operations based on tensors. An AB21 supercomputing AI processor has been proposed to accelerate such applications. This study proposes a XEM accelerator to accelerate linear algebraic operations in an AB21 processor effectively. The XEM accelerator has outer product-based parallel floating-point units that can efficiently process tensor operations. We provide hardware details of the XEM architecture and introduce new instructions for controlling the XEM accelerator. Additionally, hardware characteristic analyses based on chip fabrication and simulator-based functional verification are conducted. In the future, the performance and functionalities of the XEM accelerator will be verified using an AB21 processor.

The photochemical acid generation is refined from the first principles of quantum electrodynamics. First, we briefly review the formulation of the quantum theory of light based on the quantum electrodynamics framework to establish the probability of acid generation at a given spacetime point. The quantum mechanical acid generation is then combined with the deprotection mechanism to obtain a probabilistic description of the deprotection density directly related to feature formation in a photoresist. A statistical analysis of the random deprotection density is presented to reveal the leading characteristics of stochastic feature formation.

This paper presents a membrane computation error-minimized mixed-mode spiking neural network (SNN) crossbar array. Our approach involves implementing an embedded dummy switch scheme and a mid-node pre-charge scheme to construct a high-precision current-mode synapse. We effectively suppressed charge sharing between membrane capacitors and the parasitic capacitance of synapses that results in membrane computation error. A 400 × 20 SNN crossbar prototype chip is fabricated via a 28-nm FDSOI CMOS process, and 20 MNIST patterns with their sizes reduced to 20 × 20 pixels are successfully recognized under 411 μW of power consumed. Moreover, the peak-to-peak deviation of the normalized output spike count measured from the 21 fabricated SNN prototype chips is within 16.5% from the ideal value, including sample-wise random variations.

This paper presents the basic quantum reinforcement learning theory and its applications to various engineering problems. With the advances in quantum computing and deep learning technologies, various research works have focused on quantum deep learning and quantum machine learning. In this paper, quantum neural network (QNN)-based reinforcement learning (RL) models are discussed and introduced. Moreover, the pros of the QNN-based RL algorithms and models, such as fast training, high scalability, and efficient learning parameter utilization, are presented along with various research results. In addition, one of the well-known multi-agent extensions of QNN-based RL models, the quantum centralized-critic and multiple-actor network, is also discussed and its applications to multi-agent cooperation and coordination are introduced. Finally, the applications and future research directions are introduced and discussed in terms of federated learning, split learning, autonomous control, and quantum deep learning software testing.

The point cloud provides viewers with intuitive geometric understanding but requires a huge amount of data. Moving Picture Experts Group (MPEG) has developed video-based point-cloud compression in the range of 300–700. As the compression rate increases, the complexity increases to the extent that it takes 101.36 s to compress one frame in an experimental environment using a personal computer. To realize real-time point-cloud compression processing, the direct patch projection (DPP) method proposed herein simplifies the complex patch segmentation process by classifying and projecting points according to their geometric positions. The DPP method decreases the complexity of the patch segmentation from 25.75 s to 0.10 s per frame, and the entire process becomes 8.76 times faster than the conventional one. Consequently, this proposed DPP method yields similar peak signal-to-noise ratio (PSNR) outcomes to those of the conventional method at reduced times (4.7–5.5 times) at the cost of bitrate overhead. The objective and subjective results show that the proposed DPP method can be considered when low-complexity requirements are required in lightweight device environments.

In wireless sensor network (WSN) monitoring systems, redundant data from sluggish environmental changes and overlapping sensing ranges can increase the volume of data sent by nodes, degrade the efficiency of information collection, and lead to the death of sensor nodes. To reduce the energy consumption of sensor nodes and prolong the life of WSNs, this study proposes a dual layer intracluster data fusion scheme based on ring buffer. To reduce redundant data and temporary anomalous data while guaranteeing the temporal coherence of data, the source nodes employ a binarized similarity function and sliding quartile detection based on the ring buffer. Based on the improved support degree function of weighted Pearson distance, the cluster head node performs a weighted fusion on the data received from the source nodes. Experimental results reveal that the scheme proposed in this study has clear advantages in three aspects: the number of remaining nodes, residual energy, and the number of packets transmitted. The data fusion of the proposed scheme is confined to the data fusion of the same attribute environment parameters.

Author verification is challenging because of the diversity in writing styles. We propose an enhanced handwriting verification method that combines handcrafted and automatically extracted features. The method uses a genetic algorithm to reduce the dimensionality of the feature set. We consider offline Bangla handwriting content and evaluate the proposed method using handcrafted features with a simple logistic regression, radial basis function network, and sequential minimal optimization as well as automatically extracted features using a convolutional neural network. The handcrafted features outperform the automatically extracted ones, achieving an average verification accuracy of 94.54% for 100 writers. The handcrafted features include Radon transform, histogram of oriented gradients, local phase quantization, and local binary patterns from interwriter and intrawriter content. The genetic algorithm reduces the feature dimensionality and selects salient features using a support vector machine. The top five experimental results are obtained from the optimal feature set selected using a consensus strategy. Comparisons with other methods and features confirm the satisfactory results.

Music identification is widely regarded as a solved problem for music searching in quiet environments, but its performance tends to degrade in TV broadcast audio owing to the presence of dialogue or sound effects. In addition, constructing an accurate dataset for measuring the performance of background music monitoring in TV broadcast audio is challenging. We propose a framework for monitoring background music by automatic identification and introduce a background music cue sheet. The framework comprises three main components: music identification, music–speech separation, and music detection. In addition, we introduce the Cue-K-Drama dataset, which includes reference songs, audio tracks from 60 episodes of five Korean TV drama series, and corresponding cue sheets that provide the start and end timestamps of background music. Experimental results on the constructed and existing datasets demonstrate that the proposed framework, which incorporates music identification with music–speech separation and music detection, effectively enhances TV broadcast audio monitoring.