Search

Existing tracking algorithms mostly rely on model-driven approaches, which can be prone to inaccuracies due to unpredictable human behaviours. This article aims to address the issue of transient errors in tracking port container trucks (PCTrucks) when encountering obstructions. A data-driven algorithm for predicting vehicle trajectories is proposed in this study. The approach involves preprocessing an extensive dataset of GPS information, training a DeepLSTM-Attention model, and integrating the proposed model with the population-based training (PBT) algorithm to optimise network hyperparameters. The objective is to enhance the accuracy of predicting trajectories for vehicles moving horizontally. The trajectory data used are collected from real-world port operations. This research is conducted across nine trajectory segments and benchmarked against traditional approaches like Kalman filtering, machine learning techniques such as support vector regression (SVR) and standard long short-term memory (LSTM) networks. The results demonstrate that the proposed prediction method, that is, DeepPBM-Attention, outperforms other techniques in several evaluation metrics, including root mean square error (RMSE), mean absolute error (MAE), F1 score and trajectory reconstruction error (TRE). Compared to LSTM networks, the performance of DeepPBM-Attention is improved by approximately 40%. The proposed data-driven trajectory prediction algorithm exhibits high accuracy and practicality, which can effectively be applied to the positioning prediction of horizontally moving vehicles in port environments.

Automated machine learning (AutoML) is recognized for its efficiency in facilitating model development due to its ability to perform tasks autonomously, without constant human intervention. AutoML automates the development and optimization of machine learning models, leading to high energy consumption due to the large amount of calculations involved. Hyperparameter optimization algorithms, central to AutoML, can significantly impact its carbon footprint. This work introduces and investigates energy efficiency metrics for advanced hyperparameter optimization algorithms within AutoML. These metrics enable the evaluation and optimization of an algorithm's energy consumption, considering accuracy, sustainability, and reduced environmental impact. The experimentation demonstrates the application of Green AI principles to AutoML hyperparameter optimization algorithms. It assesses the current sustainability of AutoML practices and proposes strategies to make them more environmentally friendly. The findings indicate a reduction of 28.7% in CO2e emissions when implementing the Green AI strategy, compared to the Red AI strategy. This improvement in sustainability is achieved with a minimal decrease of 0.51% in validation accuracy. This study emphasizes the importance of continuing to investigate sustainability throughout the life cycle of AI, aligning with the three fundamental pillars of sustainable development. [ABSTRACT FROM AUTHOR]

Unsupervised person re-identification has achieved great success through the self-improvement of individual neural networks. However, limited by the lack of diversity of discriminant information, a single network has difficulty learning sufficient discrimination ability by itself under unsupervised conditions. To address this limit, we develop a population-based evolutionary gaming (PEG) framework in which a population of diverse neural networks are trained concurrently through selection, reproduction, mutation, and population mutual learning iteratively. Specifically, the selection of networks to preserve is modeled as a cooperative game and solved by the best-response dynamics, then the reproduction and mutation are implemented by cloning and fluctuating hyper-parameters of networks to learn more diversity, and population mutual learning improves the discrimination of networks by knowledge distillation from each other within the population. In addition, we propose a cross-reference scatter (CRS) to approximately evaluate re-ID models without labeled samples and adopt it as the criterion of network selection in PEG. CRS measures a model's performance by indirectly estimating the accuracy of its predicted pseudo-labels according to the cohesion and separation of the feature space. Extensive experiments demonstrate that (1) CRS approximately measures the performance of models without labeled samples; (2) and PEG produces new state-of-the-art accuracy for person re-identification, indicating the great potential of population-based network cooperative training for unsupervised learning. The code is released on github.com/YunpengZhai/PEG. [ABSTRACT FROM AUTHOR]

Speech is an effective, quick, and important way for communicating and exchanging complex information between humans. Emotions have always been a part of normal human conversation which makes the speech more attractive. Because of this major role of both speech and emotion, many researchers are inspired by studying Speech Emotion Recognition (SER) which still has plenty of challenges. In this study, we proposed a novel reservoir computing approach with the initialization of random connection weights for the input weight by the truncated normal distribution. Furthermore, Population-Based Training (PBT) is adopted to optimize the hyperparameters of the whole Echo State Network (ESN) model which have a significant impact on the model performance. The proposed model has adopted bidirectional reservoir input to increase the memorization capability, and Sparse Random Projection (SRP) was applied for dimensional reduction as a simple, unsupervised, and low complexity approach. The speaker-independent strategy was employed on EMODB and SAVEE datasets as an acted speech emotion dataset and Aibo as a non-acted dataset. The model achieved 84.8%, 65.95%, and 45.99% unweighted average recalls on the EMODB, SAVEE, and Aibo datasets respectively. The results show that the proposed model outperforms the recent state-of-the-art studies with a cheaper computational cost. [ABSTRACT FROM AUTHOR]

• An improved model DWTimesNet based on dilated convolution and WN is proposed. • RF was used to reduce the effect of missing data on the proposed model. • To eliminate the effect of some variables, MIC is used for feature selection. • The PBT algorithm is used to efficiently parameterize the proposed model. • Two cases with different data quality are used to verify the proposed model. The share of wind power in global electricity generation is increasing year by year, and the prediction of wind power is a practical and necessary scientific research. In this paper, the TimesNet model is used for multistep prediction of wind power. It is observed that the prediction error of the TimesNet model could be improved by replacing the original convolutional structure with the dilated convolution and weight normalization (WN). And in addition, the population-based training (PBT) algorithm is introduced to guide the model tuning. Missing values in the raw data are imputed to enhance data integrity. And the maximal information coefficient (MIC) method is chosen to select features for the different influences on wind power, which reduces the computational effort of the model. In order to demonstrate the effectiveness of the proposed model MIC-PBT-DWTimesNet, nine models are arranged in this paper as a control group and compared with their predicted values with each other. It can be observed that the RMSE of the proposed model MIC-PBT-DWTimesNet is decreased by 17–33% compared to the base TimesNet method. [ABSTRACT FROM AUTHOR]