Previous work has recognized the importance of using the attention mechanism to obtain the interaction between aspect words and contexts for sentiment analysis. However, for the most attention mechanisms, it is unrigorous to use the average vector of the aspect words to calculate the context attention. Besides, the feature extraction ability of the model is also essential for effective analysis, the combination of CNN and LSTM can enhance the feature extraction ability and semantic expression ability of the model, which is also a popular research trend. This paper introduces an aspect level neural network for sentiment analysis named Feature Enhanced Attention CNN-BiLSTM (FEA-NN). Our method is to extract a higher-level phrase representation sequence from the embedding layer by using CNN, which provides effective support for subsequent coding tasks. In order to improve the quality of context encoding and preserve semantic information, we use BiLSTM to capture both local features of phrases as well as global and temporal sentence semantics. Besides, we add an attention mechanism to model interaction relationships between aspect words and sentences to focus on those keywords of targets to learn more effective context representation. We evaluate the proposed model on three datasets: Restaurant, Laptop, and Twitter. Extensive experiments show that the effectivess of FEA-NN.

As a core task of natural language processing and information retrieval, automatic text summarization is widely applied in many fields. There are two existing methods for text summarization task at present: abstractive and extractive. On this basis we propose a novel hybrid model of extractive-abstractive to combine BERT (Bidirectional Encoder Representations from Transformers) word embedding with reinforcement learning. Firstly, we convert the human-written abstractive summaries to the ground truth labels. Secondly, we use BERT word embedding as text representation and pre-train two sub-models respectively. Finally, the extraction network and the abstraction network are bridged by reinforcement learning. To verify the performance of the model, we compare it with the current popular automatic text summary model on the CNN/Daily Mail dataset, and use the ROUGE (Recall-Oriented Understudy for Gisting Evaluation) metrics as the evaluation method. Extensive experimental results show that the accuracy of the model is improved obviously.

Text classification is a pivotal task in NLP (Natural Language Processing), which has received widespread attention recently. Most of the existing methods leverage the power of deep learning to improve the performance of models. However, these models ignore the interaction information between all the sentences in a text when generating the current text representation, which results in a partial semantics loss. Labels play a central role in text classification. And the attention learned from text-label in the joint space of labels and words is not leveraged, leaving enough room for further improvement. In this paper, we propose a text classification method based on Self-Interaction attention mechanism and label embedding. Firstly, our method introduce BERT (Bidirectional Encoder Representation from Transformers) to extract text features. Then Self-Interaction attention mechanism is employed to obtain text representations containing more comprehensive semantics. Moreover, we focus on the embedding of labels and words in the joint space to achieve the dual-label embedding, which further leverages the attention learned from text-label. Finally, the texts are classified by the classifier according to the weighted labels representations. The experimental results show that our method outperforms other state-of-the-art methods in terms of classification accuracy.

The pursuit of higher recognition accuracy and speed with smaller model sizes has been a major research topic in the detection of surface defects in steel. In this paper, we propose an improved high-speed and high-precision Efficient Fusion Coordination network (EFC-YOLO) without increasing the model’s size. Since modifications to enhance feature extraction in shallow networks tend to affect the speed of model inference, in order to simultaneously ensure the accuracy and speed of detection, we add the improved Fusion-Faster module to the backbone network of YOLOv7. Partial Convolution (PConv) serves as the basic operator of the module, which strengthens the feature-extraction ability of shallow networks while maintaining speed. Additionally, we incorporate the Shortcut Coordinate Attention (SCA) mechanism to better capture the location information dependency, considering both lightweight design and accuracy. The de-weighted Bi-directional Feature Pyramid Network (BiFPN) structure used in the neck part of the network improves the original Path Aggregation Network (PANet)-like structure by adding step branches and reducing computations, achieving better feature fusion. In the experiments conducted on the NEU-DET dataset, the final model achieved an 85.9% mAP and decreased the GFLOPs by 60%, effectively balancing the model’s size with the accuracy and speed of detection.

Multimodal Sentiment Analysis (MSA) is the technology of intelligently recognizing and assessing human sentiments using various data forms such as text, image, and audio. Despite current mainstream methods have made significant progress, MSA still faces the following issues: 1) most current methods train models based on pre-extracted features, lacking a sufficient understanding of sentiment diversity in multimodal data and may even lead to the loss of critical information in the raw data; and 2) textual modality, which possesses high-level semantic features, should typically dominate the fusion process, yet current methods fail to fully leverage this characteristic to balance modality information. To address the aforementioned issues, we propose a novel Multimodal Sentiment Analysis framework using Multimodal-Prefixed and Cross-Modal Attention (DB-MPCA). For the first issue, DB-MPCA employs multimodal raw data for pre-training, which not only allows for in-depth exploration of multimodal information but also significantly enhances the model's learning capabilities and generalization, while reducing the substantial costs associated with manual annotation. Regarding the second issue, DB-MPCA introduces two prefix encoders designed to convert acoustic and visual features into prefix tokens. These tokens are then embedded into a pre-trained language model, where they are encoded together with textual tokens. Through this approach, DB-MPCA effectively learns cross-modal attention while maintaining the dominance of the textual modality, thereby optimizing the fusion of modalities. Comprehensive experiments conducted on the widely utilized dataset(CMU-MOSI)demonstrate the effectiveness of our model, highlighting its superiority over baseline models