KalmanNet demonstrates potential for nonlinear state estimation, but it directly concatenates heterogeneous components with disparate semantics into a composite vector. This approach ignores capturing semantic component correlations, limiting the network's ability to utilize this information when computing the Kalman gain. To address this, we propose CNPE-KalmanNet, a novel learning-aided Kalman filter that models semantic component correlations by integrating element-wise rotary encoding with multilayer perceptrons in the complex domain. CNPE serves as a plug-and-play module compatible with two baseline KalmanNet architectures. Experiments on the Michigan NCLT dataset demonstrate that CNPE-KalmanNet significantly improves estimation accuracy. Compared to the existing KalmanNet, our method reduces the RMSE by 20%-25% in state estimation tasks and by 15%-20% in sensor fusion tasks.

Graph representation learning is the foundation for various graph data mining tasks. In the real world, graph data not only contains complex adjacency relationships but also diverse structural information. To address issues such as overfitting and overemphasis on neighboring information while neglecting structural information in graph autoencoders, a novel approach that combines generative learning and masked autoencoder for graph representation learning is proposed. This method employs a masked autoencoder to mask a portion of the graph structure, using the remaining structure as input to the graph autoencoder, effectively alleviating overfitting. Additionally, leveraging generative learning theory, a new graph autoencoder is introduced, capable of aggregating both neighbor and structural information to generate high-quality graph embeddings. Comparative experiments between GLMAE and representative graph representation learning methods demonstrate that GLMAE achieves state-of-the-art performance in link prediction and node classification tasks.

This article presents a novel transformer-less single-stage buck-boost inverter (SSBBI) utilizing coupled inductors and its dual-mode time-sharing control method. The leakage current can be completely suppressed by connecting the neutral of the electrolytic capacitors to the common ground of the power grid. Besides, only one MOSFET operates at high-frequency in Buck and Boost mode. As a result, in theory, high conversion efficiency can be guaranteed for the proposed inverter. Moreover, compared with conventional Aalborg inverters, the number of switches, diodes, and inductors is further reduced to reduce cost and increase power density. In addition, the operating principle based on equivalent circuits and the control strategy of the proposed inverter are given in this article. As a proof of concept, the simulation results based on PSIM of the proposed SSBBI with different input voltages and output powers and a 110 V power grid are provided.

The extended Kalman filter has been widely used in sensor fusion to achieve integrated navigation and localization. Efficiently integrating multiple sensors requires prior knowledge about their errors for setting the filter. The recently emerged KalmanNet managed to use recurrent neural networks to learn prior knowledge from data and carry out state estimation for problems under non-linear dynamics with partial information. In this letter, the KalmanNet is implemented for integrated navigation using data from GPS/Wheels and the Inertial Measurement Unit. Therein, a practical strategy for the training algorithm of truncated backpropagation through time is presented by taking advantage of the first-order Markov property of the system state of the Kalman filter, which improves the training robustness and performance of the existing KalmanNet. Experimental results on the Michigan NCLT dataset show that our fusion KalmanNet significantly outperforms the conventional EKF-based fusion algorithm with an improvement of 20%$\sim$40% in average RMSE.

To address the issues of high human interference and low efficiency in traditional manual methods for classifying and rating steel scrap, we propose the development of CSBFNet, a deep learning-based model for multi-category steel scrap classification and rating. Firstly, we built a 1:3 physical model of steel scrap quality inspection to simulate the unloading of a truck. We used a high-resolution vision sensor to capture the morphological characteristics of various steel scraps. Next, we trained the CSBFNet model using this data to obtain characteristic information for classifying and judging various types of scrap steel. Finally, we tested and improved the CSBFNet model at a Chinese steel mill. The results demonstrate that the model can effectively determine the automatic rating for different grades of scrap. The average accuracy rate of all types of steel scrap reaches 92.4% for the full category, with an mAP of 90.7%. Compared to traditional artificial quality detection methods, it has clear advantages in accuracy and fairness. This model solves the problem of evaluating the quality of steel scrap in the recycling process.

To solve the problem of poor classification performance of multiple complex emotions in acoustic modalities, we propose a hierarchical grained and feature model (HGFM). The frame-level and utterance-level structures of acoustic samples are processed by the recurrent neural network. The model includes a frame-level representation module with before and after information, a utterance-level representation module with context information, and a different level acoustic feature fusion module. We take the output of frame-level structure as the input of utterance-level structure and extract the acoustic features of these two levels respectively for effective and complementary fusion. Experiments show that the proposed HGFM has better accuracy and robustness. By this method, we achieve the state-of-the-art performance on IEMOCAP and MELD datasets.

With the prevalent of short texts, discovering the topics within them has become an important task. Biterm Topic Model (BTM) is more suitable to discover topics on short texts than traditional topic models. However, there are still some challenges that dealing short texts with BTM will always ignor e the document-topic semantic information and lack the true intentions of users. In addition, it is a static method and can not manage streaming short texts when a new one arrives immediately. In order to keep document-topic information and get the topic distribution of a new short text at once, we propose a joint model based on online algorithms of Latent Dirichlet Allocation (LDA) and BTM, which combines the merits of both models. Not only does it alleviate the sparsity when addressing short texts with the online algorithm of BTM, namely Incremental Biterm Topic Model (IBTM), but also keeps document-topic information with extended LDA. And considering the differences between English and Chinese text in writing, we use combined words in short texts as key words to extend the length of short texts and keep the true intensions of users. As shown in the experiment results on two real world datasets, our method is better than other baseline methods. In the end, we explain an application of our method in the task of discovering user interest tags.

Short texts are prevalent on today’s Web, especially with the emergence of social media. However, how to discover the topics of streaming short texts has become an important task for many content analysis applications. Conventional topic models such as Probabilistic Latent Semantic Analysis (PLSA) and Latent Dirichlet Allocation (LDA) will suffer from sparsity problem when we infer the latent topics from short texts with them. The reason is that they derive topics from document-level word co-occurrence by modeling each document as a mixture of topics. Different from the above idea, Biterm Topic Model (BTM) discovers topics in short texts by directly modeling the generation of word co-occurrence patterns in the whole corpus. But semantic information is lacking for short texts. In this paper, in order to alleviate the sparsity problem, keep the semantic information of documents and get the latent topic information of streaming short texts immediately, we propose a joint topic model for Chinese streaming short texts (CTM) based on the online algorithms of LDA and BTM. Experiments on short texts from Sina Weibo show that our joint topic model can discover more precise topics and carry out more applications. In addition, considering the preprocessing in Chinese text is different from English and errors in extracting key phrases, we use a combined word method to extend the length of short texts and reduce errors in extracting key phrases.