Abstract ：

This paper presents a deep learning-based topology optimization method for the joint design of material layout and fiber orientation in continuous fiber-reinforced composite structure (CFRCS). The proposed method mainly includes three steps: (1) a ResUNet-involved generative and adversarial network (ResUNet-GAN) is developed to establish the end-to-end mapping from structural design parameters to fiber-reinforced composite optimized structure, and a fiber orientation chromatogram is presented to represent continuous fiber angles; (2) to avoid the local optimum problem, the independent continuous mapping method (ICM method) considering the improved principal stress orientation interpolated continuous fiber angle optimization (PSO-CFAO) strategy is utilized to construct CFRCS topology optimization dataset; (3) the well-trained ResUNet-GAN is deployed to design the optimal structural material distribution together with the corresponding continuous fiber orientations. Numerical simulations for benchmark structure verify that the proposed method greatly improves the design efficiency of CFRCS along with high design accuracy. Furthermore, the CFRCS topology configuration designed by ResUNet-GAN is fabricated by additive manufacturing. Compression experiments of the specimens show that both the stiffness structure and peak load of the CFRCS topology configuration designed by the proposed method have significantly enhanced. The proposed deep learning-based topology optimization method will provide great flexibility in CFRCS for engineering applications. © The Chinese Society of Theoretical and Applied Mechanics and Springer-Verlag GmbH Germany, part of Springer Nature 2024.

Keyword ：

High modulus textile fibers Shape optimization Structural optimization Benchmarking Topology Structural dynamics Composite structures

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Action segmentation is vital for video understanding because it heuristically divides complex untrimmed videos into short semantic clips. Real-world human actions exhibit complex temporal dynamics, encompassing variations in duration, rhythm, and range of motions, etc. While deep networks have been successfully applied to these tasks, they face challenges in effectively adapting to these complex variations due to the inherent difficulty in capturing semantic information from a global perspective. Merely relying on distinguishing visual representations in local regions leads to the issue of over-segmentation. In an attempt to address this practical issue, we propose a novel approach named ASGSA, which aims to obtain smoother segmentation results by extracting instructive semantic information. Our core component, Global Semantic-Aware module, provides an effective way to encode the long-range temporal relation in the long untrimmed video. Specifically, we exploit a hierarchical temporal context aggregation, which is identified by a gated-mechanism selection to control the information passage at different scales. In addition, an adaptive fusion strategy is designed to guide the segmentation with the extracted semantic information. Simultaneously, to obtain higher-quality video representation without extra annotations, we resort to self-supervised training strategy and propose the Video Speed Prediction module. Extensive experiments demonstrate that our approach achieves state-of-the-art performance on all three challenging benchmark datasets (Breakfast, 50Salads, GTEA) and significantly improves the F1 score@50, which represents the reduction of over-segmentation. The code is available at https://github.com/ten000/ASGSA. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2024.

Keyword ：

Benchmarking Semantic Segmentation Semantic Web Semantics Complex networks

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Pedestrian trajectory prediction has become a crucial task in the field of autonomous driving. To improve the accuracy of pedestrian trajectory prediction, researchers primarily concentrate on tackling two key challenges. One is to extract the intricate interactions between pedestrians, and the other is to simulate the diverse decision-making intentions displayed by pedestrians. However, most existing methods utilize the distance attribute to build the relationship of pedestrians only, but ignore other features such as the steering. Besides, some generation theory based methods would lead to substantial deviations in the generated trajectory distribution since they always refine the variational likelihood lower bound of observed data. In this paper, we adopt Graph theory and propose a Spatial-Temporal Dual Graph neural network for pedestrian trajectory prediction. In the proposed method, we construct a pedestrian graph structure by utilizing pedestrian distance and steering features to extract more comprehensive interaction information. Additionally, we introduces the flow-based Glow-PN module to predict multi-modal trajectories of pedestrians. Experimental results on two public benchmark datasets show that our model achieves superior prediction performance and operates effectively in diverse scenarios. © 2024 IEEE.

Keyword ：

Forecasting Flow graphs Benchmarking Decision making Trajectories Graph neural networks Deep neural networks

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Wastewater treatment processes (WWTPs) are complex industrial processes with disturbance and strong nonlinearity, and it is difficult to accurately track the pre-designed dissolved oxygen concentration (DOC) in a finite time. To solve these problems, a robust dynamic surface control strategy with fixed time observer (DSRC-FTO) is proposed to achieve a stable control performance for DOC in this paper. The contributions of DSRC-FTO are three folds. First, an adaptive interval type-2 fuzzy neural network based on predictor (P-AIT2FNN) is applied to adaptively imitate the strong nonlinearity of WWTPs. Subsequently, a weight adaptive law is formulated using the predictor error to minimize modeling errors. Second, a fixed time observer (FTO), based on dynamic surface technique, is developed to actively suppress the unknown disturbances. Third, it is proved that the designed FTO can converge in fixed time. Then, the finite time stability of DSRC-FTO can be proved. Finally, DSRC-FTO is tested on the benchmark simulation model no. 1 (BSM1). Compared other existing methods, DSRC-FTO can realize accurate control for DOC in a finite time. © 2024 Technical Committee on Control Theory, Chinese Association of Automation.

Keyword ：

Control nonlinearities Benchmarking Fuzzy inference Wastewater treatment Robust control Fuzzy neural networks Robustness (control systems)

Cite：

Copy from the list or Export to your reference management。

Abstract ：

In recent years, there has been a substantial increase in the amount of visual data generated by edge devices. Machines typically process this data to accomplish tasks such as object detection without human visual judgment. However, human viewing is sometimes required during human-robot interaction. Here, there exists a significant difference in the focus of information between humans and machines. To tackle this issue, we propose an end-to-end learning-based image coding framework, aiming to strike a balance between human and machine vision tasks. Also, a portion of the latent space is used for both machine vision and human vision. This is different from a compression framework that only targets human vision. Because of this difference, correlations still exist between tasks. So we propose a partial-channel context model to improve coding performance.Our scalable coding framework achieves simultaneous support for both human and machine vision by partitioning the latent space. Machine vision tasks are handled by a subset of the latent space, referred to as the base layer. More complex human visual reconstruction tasks are accomplished by an additional subset of the latent space, comprising both base and enhancement layers. In the experimental section, we present the performance of human visual reconstruction and machine vision tasks, comparing them with other benchmarks. The experiments demonstrate that our framework achieves a 28.27%-38.16% reduction in bitrate for machine vision tasks and matches the performance of state-of-the-art image codecs in terms of input reconstruction. © 2024 IEEE.

Keyword ：

Benchmarking Human robot interaction Image coding Deep neural networks Video signal processing Computer vision Object detection Image compression

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Imitation learning algorithms for robotics applications require sufficient optimal data to learn well-performing strategies. State-of-the-art approaches utilize pre-labeled data or interaction with the environment to filter suboptimal data, which is time-consuming and laborious in reality. In this paper, we propose a new approach that avoids manual labeling or environment interaction. We design an additional discriminator for the behavioral cloning approach to distinguish the optimal and suboptimal data in order to influence policy learning and avoid suboptimal behaviors. Within this framework, we design a new imitation learning algorithm that utilizes the output of the discriminator as weights to learn efficiently on datasets containing suboptimal data. We evaluate the performance of the proposed method in four environments and compare it with three benchmark methods. The results illustrate that our method has better performance when dealing with datasets containing suboptimal data. The method we proposed can distinguish data with higher values in the dataset and enable the agent to learn high-performance policy from imperfect demonstrations or a small amount of data. © 2024 IEEE.

Keyword ：

Benchmarking Discriminators Learning algorithms Clone cells Cloning

Cite：

Copy from the list or Export to your reference management。

Abstract ：

With the rapid development of the Internet, various network invasive behaviors are increasing rapidly. This seriously threatens the economic development of individuals, enterprises, and society. Network intrusion detection is important in network security systems, which can be regarded as a classification problem. It aims to distinguish between the specific categories of various network behaviors and determine whether the behavior belongs to network intrusion. However, network intrusions present a diverse and fast-changing trend, making categorizing difficult. Due to feature redundancy, uneven distribution of sample numbers, and inefficient parameter optimization, traditional rule-based approaches fail to achieve satisfying classification accuracy. This work proposes a multi-classification intrusion detection model based on Stacked Sparse Shrink AutoEncoder (SSSAE), Genetic Simulated annealing-based particle swarm optimization optimized Tabnet classifier (GS-Tabnet), and Decision Fusion (DF), called for SGTD short. Among them, SSSAE extracts multiple feature sets from the input data. Then GS-Tabnet trains a classifier for each feature set. Finally, the decision fusion fuses the results from these classifiers to obtain the final classification result. SGTD is compared with eight multi-classification benchmark models, and its intrusion detection accuracy is superior to its peers. © 2024 IEEE.

Keyword ：

Simulated annealing Particle swarm optimization (PSO) Network security Network intrusion Genetic algorithms Benchmarking Intrusion detection

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Deep learning (DL) has emerged as a powerful tool for predicting the remaining useful lifetime (RUL) of components and systems. However, there are two challenges. First, DL-based methods require a sufficient number of labelled samples, while the number of run-to-failure units in practical is often small due to the high cost and time-consuming life test. Second, it is desirable to provide the prediction uncertainty of RUL for maintenance decision-making. To tackle above issues, we propose a novel few-shot RUL prediction model with a hypernetwork structure incorporating uncertainty quantification and calibration (Hyper-UC). The proposed Hyper-UC uses a shared feature embedding network and a unit-specific weight to model the mapping from sensor signals to RUL, and a generative network is designed to learn the meta-knowledge of generating distributions over the unit-specific weights. To accurately provide the predictive uncertainty, the Hyper-UC systematically models two types of uncertainty: epistemic uncertainty and aleatoric uncertainty. In the online phase, the unit-specific weights of in-service units are obtained through calibration, and subsequently the predictive distribution of the RUL can be obtained. The proposed method is evaluated to have superior model performance than the benchmark methods in a case study using the C-MAPSS dataset. © 2024 IEEE.

Keyword ：

Network embeddings Prediction models Zero-shot learning Generative adversarial networks Benchmarking Deep learning Uranium compounds Decision making

Cite：

Copy from the list or Export to your reference management。

Abstract ：

In the practical industrial field, optimizing parameter configurations during factory operations is an indispensable step, especially for multi-objective optimization problems (MOPs) that involve high computational, economic costs or extensive numerical simulations, which was categorized as Expensive Multi-Objective Optimization Problems (EMOPs). To efficiently and reasonably evaluate fitness in the face of EMOP, it is essential to develop a surrogate model to minimize costly expensive function evaluations and assist the entire optimization process. This study proposes a Surrogate-Assisted Evolutionary Algorithm SFEA that integrates Support Vector Machines (SVM) and Feedforward Neural Networks (FNN). SFEA constructs surrogate models using heterogeneous integration methods to mitigate the limitations of a single-model approaches in complex and variable optimization scenarios, thereby enhancing the efficiency and quality of the optimization outcomes. Additionally, this paper proposes an adaptive sampling strategy and a corresponding sample filling mechanism within a dual-archive management framework, based on the convergence and diversity indices of the algorithm. Experimental validation on multi-objective DTLZ and WFG benchmark problems confirms the effectiveness and feasibility of SFEA in addressing expensive multi-objective optimization challenges. © 2024 Technical Committee on Control Theory, Chinese Association of Automation.

Keyword ：

Feedforward neural networks Benchmarking Cost functions Records management Costs Multiobjective optimization

Cite：

Copy from the list or Export to your reference management。

Abstract ：

Increases in population and prosperity are linked to a worldwide rise in garbage. The "classification" and "recycling" of solid waste is a crucial tactic for dealing with the waste problem. This paper presents a new twolayer intelligent decision system for waste sorting based on fused features of Deep Learning (DL) models as well as a selection of an optimal deep Waste-Sorting Model (WSM) based on Multi-Criteria Decision Making (MCDM). A dataset comprising 1451 samples of images of waste, distributed across four classes - cardboard (403), glass (501), metal (410), and general trash (137), was used for sorting. This study proposes a Multi-Fused Decision Matrix (MFDM) based on identified fusion score level rules, evaluation criteria, and deep fused waste-sorting models. Five fusion rules used in the sorting process and the evaluation perspectives into the MFDM are sum, weighted sum, product, maximum, and minimum rules. Additionally, each of entropy and Visekriterijumska Optimizacija i Kompromisno Resenje in Serbian (VIKOR) methods was used for weighting selected criteria as well as ranking deep WSMs. The highest accuracy rate of 98% was scored by ResNet50-GoogleNet- Inception based on the minimum rule. However, under the same rule, an insufficient accuracy rate of sorting was presented by ResNet50-GoogleNet-Xception. Since Qi = 0 for Inception-Xception, the final output based on MCDM methods indicates that the fused Inception-Xception model outperforms the other fused deep WSMs, which achieved the lowest values of Qi. Thus, Inception-Xception was chosen as the best deep waste-sorting model based on images of waste, multiple evaluation criteria, and different fusion perspectives. The mean and standard deviation metrics were both used to validate the selection findings objectively. The suggested approach can aid urban decisionmakers in prioritizing and choosing an Artificial Intelligence (AI)-optimized optimal sorting model.

Keyword ：

Inception-xception Fusion Waste sorting Deep learning Entropy Benchmarking

Cite：

Copy from the list or Export to your reference management。