Resilient fault diagnosis and recovery framework for driving modules of electric vehicles using continuous wavelet transform-based fault-aware deep learning

Optimal hybrid fault recovery in a team of unmanned aerial vehicles

The IT asset is a core part that supports the management objective of an organization, and the fast settlement of the IT asset fault is very important. In this study, a fault recovery prediction technique is proposed, which uses the existing fault data to address the IT asset fault. The proposed fault recovery prediction technique is as follows. First, the existing fault recovery data were pre-processed and classified by fault recovery type; second, a rule was established for the keyword mapping of the classified fault recovery types and reported data; and third, a machine learning process that allows the prediction of the fault recovery method based on the established rule was presented. To verify the effectiveness of the proposed machine learning process, company A`s 33,000 computer fault data for the duration of six months were tested. The hit rate for fault recovery prediction was approximately 72%, and it increased to 81% via continuous machine learning.

Continual learning in medical devices: FDA's action plan and beyond

Unsupervised fault detection and recovery for intelligent robotic rollators

The multi-layer structures of Deep Learning facilitate the processing of higher-level abstractions from data, thus leading to improved generalization and widespread applications in diverse domains with various types of data. Each domain and data type presents its own set of challenges. Real-world time series data may have a non-stationary data distribution that may lead to Deep Learning models facing the problem of catastrophic forgetting, with the abrupt loss of previously learned knowledge. Continual learning is a paradigm of machine learning to handle situations when the stationarity of the datasets may no longer be true or required. This paper presents a systematic review of the recent Deep Learning applications of sensor time series, the need for advanced preprocessing techniques for some sensor environments, as well as the summaries of how to deploy Deep Learning in time series modeling while alleviating catastrophic forgetting with continual learning methods. The selected case studies cover a wide collection of various sensor time series applications and can illustrate how to deploy tailor-made Deep Learning, advanced preprocessing techniques, and continual learning algorithms from practical, real-world application aspects.

In this Chapter we compare 45-year-old Logos Model with AI’s deep learning technology and the neural net translation (NMT) technology that deep learning has given rise to. At a strictly computational level, Logos Model bears zero relationship to NMT, but we point out a number of ways in which Logos Model may nevertheless be seen to have anticipated NMT, specifically at the level of architecture and function. We take note of the fact that NMT has drifted away from interest in the biological verisimilitude of its models, and we note what experts say about the negative effect this has had on so-called continual machine learning (where new learning does not interfere with old learning, an obvious, vital requirement in MT). We discuss the related need for generalizations in MT learning, generalizations that are both semantic as well as syntactic, generalizations akin to the function exhibited by the brain in continual learning and processing of language. Our discussion turns on a particular point that experts at Google Deep Mind are making about continual learning, one they say that that AI has overlooked, and that, from our perspective, bears critically on MT. It concerns the way that the declarative, similarity-based operations of the hippocampus complements the more analytical, procedure-based operations of the neocortex to support continual learning. Most telling in this regard is their assertion that hippocampal learning is more than “item specific” and that, to the contrary, it exhibits distinct powers of semantic generalization. We note with satisfaction how that assertion about the complementary nature of hippocampal-neocortex learning comports with the analogical/analytical aspects of language processing in Logos Model. The very name and nature of SAL pattern-rules in Logos Model suggest this complementarity. We contend that the views of these deep learning experts provide indirect neuroscientific support for an MT methodology that affords continual, complexity-free learning, one that is predicated upon hippocampal/neocortex-like generalizations (viz., semantico-syntactic patterns). The present Chapter concludes with a Logos Model exercise illustrating the effectiveness of these declarative, hippocampal-like processes for MT.

Deep reinforcement learning has achieved significant success in various domains. However, it still faces a huge challenge when learning multiple tasks in sequence. This is because the interaction in a complex setting involves continual learning that results in the change in data distributions over time. A continual learning system should ensure that the agent acquires new knowledge without forgetting the previous one. However, catastrophic forgetting may occur as the new experience can overwrite previous experience due to limited memory size. The dual experience replay algorithm which retains previous experience is widely applied to reduce forgetting, but it cannot be applied in scalable tasks when the memory size is constrained. To alleviate the constrained by the memory size, we propose a new continual reinforcement learning algorithm called Self-generated Long-term Experience Replay (SLER). Our method is different from the standard dual experience replay algorithm, which uses short-term experience replay to retain current task experience, and the long-term experience replay retains all past tasks’ experience to achieve continual learning. In this paper, we first trained an environment sample model called Experience Replay Mode (ERM) to generate the simulated state sequence of the previous tasks for knowledge retention. Then combined the ERM with the experience of the new task to generate the simulation experience all previous tasks to alleviate forgetting. Our method can effectively decrease the requirement of memory size in multiple tasks, reinforcement learning. We show that our method in StarCraft II and the GridWorld environments performs better than the state-of-the-art deep learning method and achieve a comparable result to the dual experience replay method, which retains the experience of all the tasks.

Sentiment Analysis is a useful tool in several social and business contexts. Aspect Sentiment Classification is a subtask in Sentiment Analysis that gives information about features or aspects of people, entities, products, or services present in reviews. Different Deep Learning models have been proposed to solve Aspect Sentiment Classification focus on a specific domain such as restaurant,hotel, or laptop reviews. However, there are few proposals for creating a single model with high performance in multiple domains. The Continual Learning approach with neural networks has been used to solve aspect classification in multiple domains. However, avoid low aspect classification performance in Continual Learning is challenging. As a consequence, potential neural networkweight shifts in the learning process in different domains or datasets.In this paper, a novel Aspect Sentiment Classification approach is proposed. Our approach combines a Transformer Deep Learning technique with a Continual Learning algorithm in different domains. The input layer used is the pre‐trained model Bidirectional Encoder Representations from Transformers. The experiments show the efficacy of our proposal with 78 .

Sentiment Analysis is a useful tool in several social and business contexts. Aspect Sentiment Classification is a subtask in Sentiment Analysis that gives information about features or aspects of people, entities, products, or services present in reviews. Different Deep Learning models have been proposed to solve Aspect Sentiment Classification focus on a specific domain such as restaurant,hotel, or laptop reviews. However, there are few proposals for creating a single model with high performance in multiple domains. The Continual Learning approach with neural networks has been used to solve aspect classification in multiple domains. However, avoid low aspect classification performance in Continual Learning is challenging. As a consequence, potential neural networkweight shifts in the learning process in different domains or datasets.In this paper, a novel Aspect Sentiment Classification approach is proposed. Our approach combines a Transformer Deep Learning technique with a Continual Learning algorithm in different domains. The input layer used is the pre‐trained model Bidirectional Encoder Representations from Transformers. The experiments show the efficacy of our proposal with 78 .

Gesture recognition refers to a surface electromyography based method of gesture recognition becoming an essential element of edge computing systems to aid human-machine interface, wearable interfaces, and assistive technology systems. The latest advancements in deep learning have shown encouraging accuracy and are still constrained by frame-based processing, offline learning, excessive energy usage, and the incapacity to adjust to the constantly changing patterns of gestures. These limitations limit the use in edge devices with limited resources. To address them, SCaP-SNN (Self-adaptive Continuous and Predictive Spiking Neural Network) is presented as a memristive nano-neuromorphic model of adaptive and power efficient gesture recognition in the edge. Self-adaptive learning is described as dynamically modulating synaptic updates according to the variance of the signal, continual learning is used to retain previous knowledge about gestures while receiving incremental updates, and predictive learning can predict future gestures for a short period of time based on temporal spike dependence. The main innovation of this study is the design of a unified memristive spiking neural network incorporating self-adaptive learning, continual learning, and predictive intelligence for efficient gesture recognition at the edge of the network. The presented approach combines the spiking neural networks with memristive synapses modeling in order to support event-based computation, self-adaptive plasticity, lifelong learning without disastrous forgetting, and predictive inference. The online synaptic adaptation and consolidation processes allow to stabilize learning of new gestures without loss of the earlier learnt information The experimental results show a classification accuracy of 96.8%, F1-score of 0.965, low forgetting rate of 2.1% and a prediction accuracy of 91.4%. Power consumption analysis shows that the mean power consumption is 28 microjoules per inference, which is suitable in being used in low-power edges.

Financial institutions are required by regulation to report suspicious financial transactions related to money laundering. Therefore, they need to constantly monitor vast amounts of incoming and outgoing transactions. Given the involvement of many parties in money laundering, graph analytics is vital for effective monitoring. A particular challenge in detecting money laundering is that money launderers continuously adapt their tactics to evade detection. Hence, detection methods need constant fine‐tuning. Traditional machine learning models suffer from catastrophic forgetting when fine‐tuning the model on new data, thereby limiting their effectiveness in dynamic environments. Continual learning addresses this issue and enhances current anti‐money laundering (AML) practices by allowing models to incorporate new information while retaining prior knowledge. Research on continual graph learning for AML, however, is still scarce. In this review, we critically evaluate state‐of‐the‐art continual graph learning approaches for AML applications. We categorize methods into replay‐based, regularization‐based, and architecture‐based strategies within the graph neural network (GNN) framework, and we provide in‐depth experimental evaluations on both synthetic and real‐world AML data sets that showcase the effect of the different hyperparameters. Our analysis demonstrates that continual learning improves model adaptability and robustness in the face of extreme class imbalances and evolving fraud patterns. Finally, we outline key challenges and propose directions for future research. This article is categorized under: Data: Types and Structure > Graph and Network Data Data: Types and Structure > Streaming Data Statistical Learning and Exploratory Methods of the Data Sciences > Deep Learning

This paper describes an expert system for online fault detection and diagnosis of gas transmission networks, combining model- and knowledge-based methods. It consists of a set of hierarchically structured components which include signal processing, state observation, rule-based knowledge processing as well as an advanced user interface. In addition, an intelligent hypertext tool is integrated into the user interface to support the operator during fault recovery. The diagnosis system was tested with real measurement data from a medium sized gas distribution network. Its real-time capability and effectiveness for basic fault detection purposes was demonstrated by industrial applications.

This paper presents generalized logic-based methods for intelligent fault diagnosis in power electronic converters based on correlation between faults and basic measurements. Fault recovery is then applied based on this correlation by using necessary signals and quantities from existing measurements. The main purpose of the proposed fault diagnosis methods is for power electronic systems to survive from fault conditions that could occur in various components, to cope with the notion of a smart grid and extend their lifetime. The proposed methods are online, i.e., real-time or near-real-time, and can be applied to any power electronic system. Existing intelligent control of power electronic systems along with various short- and open-circuit faults in major power electronic components are reviewed. Two methods are established to diagnose faults and engage redundancy for fault recovery with one method using combinational logic and another using fuzzy logic. In both methods, two quantities are observed for each of the measured signals: 1) the signal's average value and 2) the signal's RMS value. Total harmonic distortion is also used in some simulations but is not required in experimental implementation. A systematic methodology to reduce the number of measured quantities while maintaining effective diagnosis is introduced. A solar PV microinverter in standalone mode is used as an example testing platform for the proposed methods. A simulation model is experimentally validated and the effect of each fault on different voltage and current measurements are observed, then both methods are tested in simulation and hardware. Results show the ability of both methods to diagnose several faults in the inverter's power stage along with their ability to engage redundancy for fault recovery.

Edge AI systems are increasingly being adopted in a wide range of application domains. These systems typically deploy Convolutional Neural Network (CNN) models on edge devices to perform inference, while relying on the cloud for model training. This is due to the high computational and memory demands of conventional model training, which exceeds the capabilities of resource-constrained edge devices running on tight power budgets. The dependency on the cloud for training is not suitable in many applications, where new objects or environmental conditions different from the ones present during training, are frequently encountered. In such applications, continual learning of new knowledge on the edge device becomes a necessity to avoid performance bottlenecks due to round-trip communication delays, network connectivity, and the available bandwidth. In this thesis, we propose Field-Programmable Gate Array (FPGA) based accelerator architecture and optimization strategies for a new paradigm of machine learning algorithms that is capable of continual learning. The proposed methods will enable edge FPGA systems to perform on-device deep continual learning for object classification. Specifically, the proposed methods aim to achieve real-time learning on-device, while providing for a high degree of scalability to learn a large number of classes. We first propose a FPGA accelerator for a Self-Organizing Neural Network (SONN), that can perform class-incremental continual learning in a streaming manner when combined with a CNN. The SONN model performs unsupervised learning from embedding features extracted from the CNN model by dynamically growing neurons and connections. We introduce design optimization strategies and runtime scheduling techniques to optimize resource usage, latency, and energy consumption. Experimental results based on Core50 dataset for continuous object recognition from video sequences, demonstrated that the proposed FPGA architecture outperforms CPU and GPU based counterparts in terms of latency and power. However, the SONN model grows proportionally to the classes learnt, which limits its scalability to learn a large number of classes efficiently. Next, we propose a FPGA accelerator for a Streaming Linear Discriminant Analysis (SLDA) model to overcome the scalability limitations of SONN. Similar to the SONN, SLDA performs continual learning from embedding features extracted from a CNN in a streaming manner. SLDA is highly scalable for learning a large number of classes as the network does not grow dynamically to accommodate new knowledge. We propose several design and runtime optimizations to minimize resource usage, latency, and energy consumption. Additionally, we introduce a new variant of SLDA and discuss the accuracy-efficiency trade-offs using popular datasets for continual learning, CoRE50 and CUB200. The results demonstrate that the proposed SLDA accelerator outperforms CPU and GPU counterparts, in terms of latency and energy efficiency. Finally, we demonstrate a full on-chip deep continual learning pipeline on FPGA, by integrating the proposed SLDA accelerator with Xilinx DPU, a programmable CNN accelerator IP. The design is implemented on a Xilinx Zynq Ultrascale+ MPSoC. In order to overcome the large performance bottleneck due to the communication overhead between the ARM processing system (PS) and programmable logic (PL), we implemented a Linux device driver that facilitates efficient memory mapping between PS-PL. The experimental results based on CoRE50 dataset show that the proposed pipeline is capable of performing continual learning at nearly the same latency as the inference pipeline, with only a marginal increase in energy consumption. Our results clearly demonstrate the viability of deploying real-time deep continual learning on edge AI systems that are equipped with FPGA accelerators.

Most of the research on human activity recognition has focused on learning a static model, considering that all the training instances are labeled and present in advance, while in streaming videos new instances continuously arrive and are not labeled. Moreover, these methods generally use application- specific hand-engineered and static feature models, which are not suitable for continuous learning. Some recent approaches on activity recognition use deep-learning-based hierarchical feature models, but the large size of these networks constrain them from being used in continuous learning scenarios. In this work, we propose a continuous activity learning framework for streaming videos by intricately tying together deep hybrid feature models and active learning. This allows us to automatically select the most suitable features and take the advantage of incoming unlabeled instances to improve the existing model incrementally. Given the segmented activities from streaming videos, we learn features in an unsupervised manner using deep hybrid networks, which have the ability to take the advantage of both the local hand-engineered features and the deep model in an efficient way. Additionally, we use active learning to train the activity classifier using a reduced amount of manually labeled instances. Retraining the models with a huge amount of accumulated examples is computationally expensive and not suitable for continuous learning. Hence, we propose a method to select the best subset of these examples to update the models incrementally. We conduct rigorous experiments on four challenging human activity datasets to demonstrate the effectiveness of our framework.