Recent Spiking Neural Networks (SNNs) works focus on an image classification task, therefore various coding techniques have been proposed to convert an image into temporal binary spikes. Among them, rate coding and di...
详细信息
ISBN:
(纸本)9781665405409
Recent Spiking Neural Networks (SNNs) works focus on an image classification task, therefore various coding techniques have been proposed to convert an image into temporal binary spikes. Among them, rate coding and direct coding are regarded as prospective candidates for building a practical SNN system as they show state-of-the-art performance on large-cale datasets. Despite their usage, there is little attention to comparing these two coding schemes in a fair manner. In this paper, we conduct a comprehensive analysis of the two codings from three perspectives: accuracy, adversarial robustness, and energy-efficiency. First, we compare the performance of two coding techniques with various architectures and datasets. Then, we measure the robustness of the coding techniques on two adversarial attack methods. Finally, we compare the energy-efficiency of two coding schemes on a digital hardware platform. Our results show that direct coding can achieve better accuracy especially for a small number of timesteps. In contrast, rate coding shows better robustness to adversarial attacks owing to the non-differentiable spike generation process. Rate coding also yields higher energy-efficiency than direct coding which requires multi-bit precision for the first layer. Our study explores the characteristics of two codings, which is an important design consideration for building SNNs(1).
The compression is required to reduce the size of data to store in computer storage as well as to transmit data over the internet with limited bandwidth. The genomic sequences DNA or RNA contain billions of nucleotide...
详细信息
ISBN:
(纸本)9781509016662
The compression is required to reduce the size of data to store in computer storage as well as to transmit data over the internet with limited bandwidth. The genomic sequences DNA or RNA contain billions of nucleotide bases (A, G, T, C) resulting large sized file to store in computer. In the previous compression algorithms the authors use direct coding technique where two bits are used to code a nucleotide base resulting compression ratio of 2 bits per byte (bpb). Some authors achieve a compression ratio less than 2 bpb after coding repeated bases differently. In this paper we proposed an improvement over direct coding technique that will compress both repeated and non-repeated sequences. The proposed algorithm provides better result as compared to existing algorithms. The existing direct coding algorithm compresses the non-repeated base (B) by prefixing a 0 followed by 2 bits code assigned for that base (i.e. 0B) whereas the repeated bases are compressed by prefixing 1 followed by 2 bits code assigned for that base (B) followed by 3 bits code to represent the number of repetitions (N) (i.e. 1BN). The non-repeated bases are coded by 3 bits and repeated bases by 6 bits but the existing algorithm is limited to compress the repeated sequence till 9 because of 3 bits coding. If the repeated sequence is greater than 9 then it requires a number of bits in multiple of 6 (i.e. 6, 12, and 18, so on). The propped algorithm compress the repeated base by making an improvement over existing one that will code a repeated sequence greater than 9 but less than 15 by prefixing 1 followed by base (B) followed by 111 followed by N (i.e. 1B111N) in 9 bits in place of 12 bits in existing algorithm hence the proposed algorithm provide better compression ratio.
Spiking neural networks (SNNs) are the third generation of neural networks that offer the advantages of low computational requirements, fast inference speed, and strong biological interpretability. This makes SNNs sui...
详细信息
Spiking neural networks (SNNs) are the third generation of neural networks that offer the advantages of low computational requirements, fast inference speed, and strong biological interpretability. This makes SNNs suitable for synthetic aperture radar (SAR) target recognition tasks, which are often constrained by limited computational power. This letter proposes SAR-TinySNN, a lightweight SNN architecture designed for SAR target recognition. Unlike existing SAR-related studies that predominantly rely on rate coding, SAR-TinySNN uses direct coding to encode SAR images, allowing for a more efficient coding method adapted to SAR images and achieving high target recognition accuracy, especially in scenarios with limited training samples. By integrating direct coding into a trainable SNN framework, SAR-TinySNN achieves competitive performance compared with traditional deep neural networks (DNNs) and deep SNNs on vehicle, aircraft, and ship SAR target recognition datasets, with faster inference times. The experimental results demonstrate the effectiveness of SAR-TinySNN for SAR target recognition.
For researchers doing qualitative research, interviews are a commonly used method. Data collected through interviews can be recorded through field notes, transcripts, or tape recordings. In the literature, there is a ...
详细信息
For researchers doing qualitative research, interviews are a commonly used method. Data collected through interviews can be recorded through field notes, transcripts, or tape recordings. In the literature, there is a debate regarding which of these recording methods should be used. There are issues of reliability, cost (time and money), loss of data, among others. Technology plays a pivotal role in this debate. Indeed, new technologies (e. g., direct coding) are often seen as potential replacements for older technologies (e. g., transcripts), which leads to a debate that is based on an evolution narrative (from field notes, to transcripts, to working from tape recordings). This article argues that a combination narrative should be considered where combination is better than substitution. Moreover, combining the advantages of field notes, transcripts, and working from tape recordings without accumulating each method's disadvantages is possible because of new technology. To support this argument, two technological tools (OneNote and SmartPen) are presented as a way to increase the effectiveness, efficiency, and economy of qualitative data management.
This paper examines phenotype and genotype mappings that are biologically inspired. These types of coding are used in evolutionary computation. direct and indirect encoding are studied. The determination of genotype a...
详细信息
ISBN:
(纸本)9783030878979;9783030878962
This paper examines phenotype and genotype mappings that are biologically inspired. These types of coding are used in evolutionary computation. direct and indirect encoding are studied. The determination of genotype and phenotype relationships and the connection to genetic algorithms, evolutionary programming and biology are examined in the light of newer advances. The NEAT and HyperNEAT algorithms are applied to the 2D Walker [41] problem of an agent learning how to walk. Results and findings are discussed, and conclusions are given. Indirect coding did not improve the situation. This paper shows that indirect coding is not useful in every situation.
Data transcription is often depicted as an essential and critical stage in qualitative research. As most researchers have experienced, it requires significant time and human resource investment. We focus on transcript...
详细信息
Data transcription is often depicted as an essential and critical stage in qualitative research. As most researchers have experienced, it requires significant time and human resource investment. We focus on transcription strategies, a topic typically missing from the methodology discourse. We explore the biases and challenges of each of the transcription strategies. By analysing 434 academic refereed papers from top journals, we underline the lack of scrutiny over the transcription process, its impact, and strategies taken to conduct it. We also interviewed some of the authors to better understand the challenges associated with transcription. This paper aims at contributing to more reflexivity on the existing strategies regarding transcription and how to increase transparency in qualitative research.
暂无评论