A Deep Learning Based Explainable Control System for Reconfigurable Networks of Edge Devices

Abstract

Edge devices that operate in real-world environments are subjected to unpredictable conditions caused by environmental forces such as wind and uneven surfaces. Since most edge systems such as autonomous vehicles exhibit dynamic properties, it is clear that reinforcement learning can be a powerful tool for improving system accuracy. Successful maintenance of the position of a vehicle in such environments has been recently achieved with the aid of Deep Reinforcement Learning (DRL) that dynamically adjusts the Reconfigurable Wireless Network (RWN) response. Deep Neural Networks (DNNs) are often seen as black boxes, as neither the acquired knowledge nor the decision rationale can be explained. This raises issues with the transparency and trustworthiness of the system because the underlying AI models are not governed by any mathematical or physics laws. In this paper, we explain the process of a DNN utilisation in an autonomous dynamic positioning system by gauging reactions of the DNN to predefined constraints and capturing the associated conditions that influence the DNN in a time series. We introduce a novel digitisation technique that reduces interesting patterns of time series data into single-digits to obtain a cross comparable view of the conditions. By analysing the clusters formed on this cross comparable view, we discovered multiple intensities of environmental conditions spanning across 44% of moderate conditions and 33% and 23% of harsh and mild conditions respectively. Our analysis showed that the proposed system can provide stable responses to uncertain conditions by predicting randomness.

[Retracted] Edge‐Enabled Heart Rate Estimation from Multisensor PPG Signals

“…In addition, many advanced techniques [29][30][31] have been applied in the feld of IoMT and edge computing, and these techniques promote the development and innovation of processing physiological signals, especially for PPG signals. For example, the technology stack of machine learning [32][33][34] has greatly improved the detection and inferring capabilities of related diseases benchmarked against physiological signals and medical data.…”

Artificial Intelligence (AI) and Machine Learning for Multimedia and Edge Information Processing

“…The XAI is an important issue which has yet to be comprehensively resolved, as traditional and deep neural network architectures do not offer the benefits of transparency which have high importance in safety-critical applications (e.g., autonomous vehicles). Some works on XAI can be found in [110,111]. The authors in [110] discussed a XAI framework and a pipeline workflow which consists of three stages: (1) Model Understanding; (2) Model Diagnosis; and (3) Model Refinement.…”

“…The authors in [110] discussed a XAI framework and a pipeline workflow which consists of three stages: (1) Model Understanding; (2) Model Diagnosis; and (3) Model Refinement. The authors in [111] give some recent work to explain the decision-making process and enhance the confidence of DNN-based solutions. In their work, the authors investigated DNN reactions towards predefined constraints and conditions for time series data.…”

“…This can serve as a comprehensive reference and point researchers and readers towards further and future works. Use Case 2: Intelligent Multimedia Processing on the Edge for Human Computer Interaction (HCI) and Health [99][100][101][102][103][104][105][106][107][108][109][110][111] The remainder of the paper is structured as follows: Section 2 provides initial discussions and background information for the edge paradigm, multimedia information processing and artificial intelligence. The next five sections provide in-depth discussions on five areas of interest.…”

An Explainable and Resilient Intrusion Detection System for Industry 5.0

“…DL-based IDS provides an efficient performance. However, these models lack explainability and interpretability, i.e., comprehending the underlying data proof of the prediction decisions for the behavior of the designed model [6]. Consequently, the decision lacks trust and their output cannot be further used to optimize the behaviour and reasoning offered by the sophisticated algorithm.…”