International Journal on Artificial Intelligence Tools

In fluid mechanics research, understanding the motion behavior of particles in a flow field is crucial for comprehending particle transport, mixing, and deposition processes. However, due to the complex interactions between particles and fluids, a single method is insufficient to accurately describe the particle motion. To tackle this problem, this study proposes an unsupervised heterogeneous domain adaptation method based on fuzzy principles, called Particle Flow Motion Analysis (PMFA). First, the data originating from both the source and target domains are preprocessed and subjected to Principal Component Analysis (PCA). Then, fuzzy rules are introduced for feature selection. Finally, the Maximum Mean Discrepancy (MMD) and Canonical Correlation Analysis (CCA) algorithms are employed to optimize the distribution disparities and correlations between the heterogeneous domains. By constructing domain adaptation tasks and comparing with five other methods, the performance of the proposed method is evaluated. The results demonstrate that the PFMA achieves an average accuracy of 89.44%, an average recall rate of 85.87%, and an $F$1 value of 89.26% across four tasks, outperforming the other five comparative methods. The proposed method holds significant importance in gaining in-depth understanding of particle motion phenomena in fluids and revealing the underlying physical mechanisms and patterns.

As an important reflection of the national economy system, stock market is closely related to the development of a country, which has received widespread attention of researchers in the economics. With the daily trading of the stock market, stock price forecasting has gradually been one of the common concerns in the economic analysis. Compared with traditional forecasting task, the stock price is interval data which can be handled by interval data regression or multi-output regression. Previous stock forecasting merely considers the stock price in homogeneous scenarios. However, the price distributions from different stocks may be heterogeneous. It is a challenging task to analyze the relationship between different stocks which follow heterogeneous distributions. In order to forecast stocks in heterogeneous scenarios, this paper introduces multi-output transfer learning into stock price forecasting. Compared with traditional regression or multi-output regression models, the multi-output transfer regression can predict opening price, closing price, highest price and lowest price of stocks and utilize source domain of a known stock to enhance the prediction of target stock price which may have limited known data in training set. The experimental results on four public market indices demonstrate the effectiveness of multi-output transfer regression for stock price forecasting.

Rolling bearings play an important role in rotating machinery. According to statistics, rolling bearings cause one-third faults of rotating machinery. Once a rolling bearing malfunctions, it may induce maintenance, affect work efficiency, or even cause the entire equipment to malfunction. Therefore, accurately determining the operating status of bearings is of great significance for maintaining the health of the rotating machinery. Most current fault detections of rolling bearing works focus on traditional anomaly detection models which assume the training set to follow the same distribution of test set. This assumption does not hold in fault detection of rolling bearings across different conditions and traditional anomaly detection models may be invalid. This paper introduces domain adaptation anomaly detection (DAAD) in the fault detection of rolling bearings to address this issue. DAAD can adapt anomaly detection across different distributions. The experiments of rolling bearing fault detection under single condition or across different condition show that DAAD is superior to most of the traditional anomaly detection models.

Music emotion recognition plays an important role in many applications such as music material library construction and music recommendation system. The current music emotion recognition mainly focuses on discrete emotions or continuous emotions under single scene. However, on the one hand, intensity is one of important aspects of emotion, which can be represented as emotion rank or ordinal class. On the other hand, there may be not enough music emotion data in training set, which needs to transfer music emotion recognition model learnt from the music data in a source domain. The distribution of existing music data in target domain may differ from target music dataset. In order to overcome these two issues, this paper proposes to utilize transfer ordinal label learning (TOLL) to estimate music emotion. Compared with the previous works, TOLL-based music emotion intensity estimation implements music intensity estimation through transferring the knowledge in the existing source domain to unknown target domain. The experiments on several datasets show that TOLL can achieve promising results for emotion intensity estimation in single scene or across different scenes.

The path planning of mobile robots helps robots to perceive environment using the information obtained from sensors and plan a route to reach the target. With the increasing difficulty of task, the environment the mobile robots face becomes more and more complex. Traditional path planning methods can no longer meet the requirements of mobile robot navigation in complex environment. Deep reinforcement learning (DRL) is introduced into robot navigation However, it may be time-consuming to train DRL model when the environment is very complex and the existing environment may differ from the unknown environment. In order to handle the robot navigation in heterogeneous environment, this paper utilizes deep transfer reinforcement learning (DTRL) for mobile robot path planning. Compared with DRL, DTRL does not require the distribution of the existing environment is the same as that of the unknown environment. Additionally, DTRL can transfer the knowledge of existing model to new scenario to reduce the training time. The simulations show that DTRL can reach higher success rate than DRL for heterogeneous environment robot navigation. By using local policy, it costs less time to train DTRL than DRL for a complex environment and DTRL can consume less navigation time.

Speech emotion analysis plays an important role in English teaching by analyzing the reading state of students. Teachers can dynamically adjust the teaching content according to the emotional feedback of students and improve the teaching quality of the school. Due to unstable student emotions and background noise, the accuracy of speech emotion recognition is constrained. Although multimodal data can alleviate the deficiency of a single modality, collecting and annotating multimodal samples requires a significant amount of resources. To resolve this issue, this paper proposes a novel multimodal sentiment analysis framework based on domain adaptive learning mechanisms to assist English teaching. We construct a novel multi-task variation autoencoder framework in which we simultaneously complete reconstruction and classification tasks. To improve speech emotion recognition performance, we introduce domain adaptive learning based on the Wasserstein distance between two variational hidden layers from the video domain (source domain) and speech domain (target domain). To validate the effectiveness of our proposed model, we conducted extensive comparative experiments on two public datasets and a self-built English oral dataset. All experimental results indicate that domain adaptation learning mechanisms can effectively improve the recognition performance of the target domain. On the self-built dataset for English teaching, the proposed model achieves higher performance compared to other deep models.

Deep networks have achieved great success in forest fire detection by exploiting visible light images. However, visible light images are susceptible to strong light, smoke, and obstruction interference. The infrared image has high sensitivity to temperature changes of targets, which can alleviate the deficiency of visible light image. Due to the significant distribution shift between visible light and infrared images, directly using the visible light-based pre-trained network for infrared forest fire results in a significant decrease in performance. To resolve this issue, this paper proposes an infrared image forest fire detection system based on domain adaptive learning. We adopt two YOLOv5 frameworks to extract features from visible light images (source domain) and infrared images (target domain). To align the features of the two domains, we construct a novel adaptation learning mechanism based on Kullback–Leibler (KL) loss and feature maximum mean discrepancy (FMMD) loss. We conducted extensive comparative experiments on two publicly available datasets to verify the effectiveness of the proposed model. All experimental results indicate that our proposed domain adaptive learning mechanism effectively improves the performance of infrared forest fire detection.

Human skeleton-based posture anomaly detection has been widely applied in the field of physical education teaching. The existing spatio-temporal graph convolutional networks (ST-GCN) can fully utilize the local and global information of the human skeleton for action recognition, but the entire model requires a large amount of computation and the modeling of high-order relationships between joint points of the human skeleton is insufficient. To this end, this paper proposes a novel domain adaptive hypergraph convolutional network for basketball posture anomaly analysis by exploiting 2D skeleton information. First, we designed an effective hypergraph convolution feature extraction network to improve the high-order dependency modeling. To further improve the performance of the hypergraph convolutional network, we introduce domain adaptive learning technology to supervise the feature extraction learning of the target domain (2D skeleton) through the source domain (3D skeleton). At last, we construct a basketball action teaching analysis dataset for model evaluation. We conducted a large number of comparative experiments on the public dataset NTU RGB+D and our self-built dataset. All the results showed that our proposed hypergraph convolutional model effectively extracts features of 2D human skeletons, and by introducing domain adaptive learning, the performance of basketball anomaly detection is further improved.

Deep learning-based models have achieved promising performance for video behavior anomaly detection, but these models require high computational complexity for processing video data. Moreover, the performance is severely restricted by different challenges, i.e., lighting conditions, background noise and occlusion. The human skeleton has a compact spatial structure and rich semantic information, which can be more robust to video data defection. Although the human skeleton has good real-time performance, the recognition accuracy still needs further improvement. To this end, we propose a novel discrete variational feature transfer learning (DVFTL) framework in which the spatiotemporal graph-embedded Transformer module is designed to construct feature extraction backbones. Specifically, we extract human skeleton information from video frames, and combine graph convolution with the Transformer encoder to explore the local and global dependencies of joint points in the human skeleton. To analyze abnormal behavior uncertainty, we constructed the extraction network based on the discrete variational reconstruction mechanism. Moreover, to further improve the skeleton detection performance, we introduce the distillation transfer learning mechanism from the video variational network to the skeleton variational network. we conducted extensive comparative experiments on two publicly available datasets. The experimental results show that the skeleton-based variational network achieves high performance. Moreover, the introduction of the transfer learning mechanism can further improve the performance of our proposed model.