Enabling resource-efficient edge intelligence with compressive sensing-based deep learning

Machidon, Alina L.; Pejović, Veljko

doi:10.1145/3528416.3530230

Cited by 6 publications

(4 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, a type of ML technique inspired by the structure and functioning of the human brain, called NN, also received considerable attention. There were 11 articles, which are [31], [32], [34], [41], [56], [65], [69], [71], [73], [81], [87], found in this study. In the context of business plans and logistics decision processes, five articles address various challenges and research problems in improving production processes and decisionmaking using AI in Industry 4.0.…”

mentioning

confidence: 65%

“…The RL class is used to provide a planning model of the mobile base agent of the SwarmItFIX robot with novel Swing and Dock (SaD) locomotion for material handling/transfer applications [54], and also combined with CV to create automated quality inspection in the production system [77]. The NN class is used as a model recommendation in the decision-making process [31], and combined with the compressive sensing (CS) method, enabling resource-efficient edge hardware intelligence [81]. Then, the SSL class is used to provide fault diagnostic and prognostic considering conditionbased maintenance (CM) data along with event-triggered data [47].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Artificial intelligence and internet of things in manufacturing decision processes

Wijaya,

Hermanto Rudy,

Debora

et al. 2024

IJ-AI

View full text Add to dashboard Cite

This paper explores the influence of the internet of things (IoT) and artificial intelligence (AI) on the decision-making processes of modern manufacturing systems. With the proliferation of IoT devices and the development of AI technologies, manufacturing companies increasingly leverage these technologies to improve their decision-making abilities. This study aims to investigate the potential benefits, difficulties, and ramifications of integrating IoT and AI in manufacturing systems. The review employs the preferred reporting items for systematic reviews and meta-analyses (PRISMA) method with a systematic mapping process with four research questions. A total of 1282 articles were collected between 2017 and 2023, reviewed in accordance with the inclusion and exclusion criteria, and 66 articles were chosen. The research on IoT and AI technologies influentially affects other research in the production control layer manufacturing area based on the top-ten cited articles. In contrast, the research in this area focused on the operations management layer, specifically manufacturing analytics processes. This paper’s findings contribute to a greater understanding of the impact of IoT and AI on decision-making in modern multi-domain manufacturing systems and provide direction for future research in this field.

show abstract

mentioning

confidence: 65%

mentioning

confidence: 99%

Artificial intelligence and internet of things in manufacturing decision processes

Wijaya,

Hermanto Rudy,

Debora

et al. 2024

IJ-AI

View full text Add to dashboard Cite

show abstract

“…As a result, there is a potential for previously prohibitively demanding continuous sensing and inference to finally be realized in certain domains. Furthermore, a graceful degradation in end-result quality can be supported with the CS-DL pipeline (Machidon and Pejović 2022). Through reduced CS sampling and reduced accuracy DL inference we can, in a controlled manner, trade result quality for resource usage.…”

Section: Towards Deep Learning-supported Compressive Sensingmentioning

confidence: 99%

“…For handling the size mismatch between the size-varying CS measurement vectors and input layer two approaches are explored: one that zero-pads the measurement vectors so they are all of the maximum length, which is also the dimension of the input layer of network; and the other that projects back into the original space dimension the measurements, to get a pseudo-inverse of the measurement matrix, like in the previous work (Lohit et al 2018b). CS rate adaptivity was also addressed in Machidon and Pejović (2022), where a zero padding strategy was also proposed and combined with context awareness to intelligently adapt the sampling rate according to the nature of the signal at the input.…”

Section: Measurement Rate Adaptivitymentioning

confidence: 99%

Deep learning for compressive sensing: a ubiquitous systems perspective

Machidon

Pejović

2022

Artif Intell Rev

Self Cite

View full text Add to dashboard Cite

Compressive sensing (CS) is a mathematically elegant tool for reducing the sensor sampling rate, potentially bringing context-awareness to a wider range of devices. Nevertheless, practical issues with the sampling and reconstruction algorithms prevent further proliferation of CS in real world domains, especially among heterogeneous ubiquitous devices. Deep learning (DL) naturally complements CS for adapting the sampling matrix, reconstructing the signal, and learning from the compressed samples. While the CS–DL integration has received substantial research interest recently, it has not yet been thoroughly surveyed, nor has any light been shed on practical issues towards bringing the CS–DL to real world implementations in the ubiquitous computing domain. In this paper we identify main possible ways in which CS and DL can interplay, extract key ideas for making CS–DL efficient, outline major trends in the CS–DL research space, and derive guidelines for the future evolution of CS–DL within the ubiquitous computing domain.

show abstract

Adaptive data-driven subsampling for efficient neural network inference

Machidon,

Pejović

2024

SIViP

View full text Add to dashboard Cite

In this paper we present a novel data-driven subsampling method that can be seamlessly integrated into any neural network architecture to identify the most informative subset of samples within the original acquisition domain for a variety of tasks that rely on deep learning inference from sampled signals. In contrast to existing methods that require signal transformation into a sparse basis, expensive signal reconstruction as an intermediate step, and that can support a single predefined sampling rate only, our approach allows the sampling inference pipeline to adapt to multiple sampling rates directly in the original signal domain. The key innovations enabling such operation are a custom subsampling layer and a novel training mechanism. Through extensive experiments with four data sets and four different network architectures, our method demonstrates a simple yet powerful sampling strategy that allows the given network to be efficiently utilized at any given sampling rate, while the inference accuracy degrades smoothly and gradually as the sampling rate is reduced. Experimental comparison with state-of-the-art sparse sensing and learning techniques demonstrates competitive inference accuracy at different sampling rates, coupled with a significant improvement in computational efficiency, and the crucial ability to operate at arbitrary sampling rates without the need for retraining.

show abstract

Enabling resource-efficient edge intelligence with compressive sensing-based deep learning

Cited by 6 publications

References 25 publications

Artificial intelligence and internet of things in manufacturing decision processes

Artificial intelligence and internet of things in manufacturing decision processes

Deep learning for compressive sensing: a ubiquitous systems perspective

Adaptive data-driven subsampling for efficient neural network inference

Contact Info

Product

Resources

About