Multi-Level Structured Self-Attentions for Distantly Supervised Relation Extraction

In this paper, a multi-instance learning framework is introduced to solve the problem of skin biopsy image features recognition. Previously reported methods for skin surface images were mostly based on color features extraction. They are incapable to be directly applied to skin biopsy image features recognition because biopsy images are often dyed and have obvious inner structures with different textures. Therefore, we regard skin biopsy images as multi-instance samples, whose instances are regions or structures captured by applying Normalized Cut. Texture feature extraction methods are used to express each region as a vectorial expression. Then two multi-instance learning algorithms reported successful in various image retrieval tasks were applied. Nine features were manually selected as target features to evaluate the proposed method on a skin disease diagnosis datasets of 6579 biopsy images from 2010 to 2011. The result showed that the proposed method is effective and medically acceptable.

Recurrent neural network (RNN) has been broadly applied to natural language process (NLP) problems. This kind of neural network is designed for modeling sequential data and has been testified to be quite efficient in sequential tagging tasks. In this paper, we propose to use bi-directional RNN with long short-term memory (LSTM) units for Chinese word segmentation, which is a crucial task for modeling Chinese sentences and articles. Classical methods focus on designing and combining hand-craft features from context, whereas bi-directional LSTM network (BLSTM) does not need any prior knowledge or pre-designing, and is expert in creating hierarchical feature representation of contextual information from both directions. Experiment result shows that our approach gets state-of-the-art performance in word segmentation on both traditional Chinese datasets and simplified Chinese datasets.

Distant supervision assumes that any sentence containing the same entity pairs reflects identical relationships. Previous works of distantly supervised relation extraction (DSRE) task generally focus on sentence-level or bag-level de-noising techniques independently, neglecting the explicit interaction with cross levels. In this paper, we propose a hierarchical contrastive learning Framework for Distantly Supervised relation extraction (HiCLRE) to reduce noisy sentences, which integrate the global structural information and local fine-grained interaction. Specifically, we propose a three-level hierarchical learning framework to interact with cross levels, generating the de-noising context-aware representations via adapting the existing multi-head self-attention, named Multi-Granularity Recontextualization. Meanwhile, pseudo positive samples are also provided in the specific level for contrastive learning via a dynamic gradient-based data augmentation strategy, named Dynamic Gradient Adversarial Perturbation. Experiments demonstrate that HiCLRE significantly outperforms strong baselines in various mainstream DSRE datasets.

ConvAttenMixer: Brain tumor detection and type classification using convolutional mixer with external and self-attention mechanisms

In this paper, a new model based on the combination of improved LSTM and self-attention mechanism is studied for the detection of nystagmus caused by vestibular illusion in pilots during flight. An efficient and robust nystagmus detection method was proposed by constructing experimental simulation scenarios and collecting and analyzing pilot eye movement data. The improved LSTM model enhances the ability of capturing the medium and long term dependence of the ocular shock sequence by adding a gating unit, and the introduction of self-attention mechanism further improves the analytical accuracy of the model for complex eye movement sequences. The experimental results show that the model has excellent performance in accuracy, recall rate and F1 score, which is significantly better than the traditional model, providing a new technical means for the detection of vestibular illusion.The LSTM-GRU-Attention model has been experimentally verified to perform best in accuracy, recall, and F1 score, reaching 095, 0.91, and 0.93 respectively, indicating that the outperforms the other two models in overall classification performance, positive sample recognition ability, and balance between accuracy and recall.

In opportunistic networks, it is difficult to find the best relay node, which not only makes the message transmission inefficient but also wastes resources in the network. Most of the existing routing algorithms based on node similarity often choose relay nodes from a single point of view such as time or space. In this paper, a new neural network architecture is designed, namely, OSAN. It combined with the self- attention mechanism, the spatial and temporal variation characteristics of nodes in two dimensions are taken into account comprehensively to the maximum. Firstly, we divide the opportunistic networks into opportunity network snapshots according to the time window and input each opportunity network snapshot into the spatial structure self-attention layer. At the same time, to capture the interaction between different snapshots, we use the temporal self-attention mechanism. Thus, the Spatio-temporal characteristics of nodes in different snapshots are extracted. Finally, the similarity between nodes is calculated according to the Spatio-temporal characteristics extracted by nodes, so we propose a Spatio-temporal topology routing algorithm in opportunistic networks based on the self-attention mechanism (STSA). The simulation results show that STSA has advantages in several common performance indexes.KeywordsOpportunistic networksDynamic graph embeddingSelf-attention mechanismNode spatio-temporal feature extractionRouting algorithm

In order to improve the accuracy of sentiment analysis, this paper presents a new method for sentiment analysis of BiLSTM-cnn text based on self attention mechanism and dense connection. Methods BiLSTM-CNN-Attmodel was established. Firstly, BiLSTM was introduced to extract context words. Then, the convolutional neural network(CNN) is used to extract local semantic features. Combined with DenseNet dense connection module, the memory strength of the whole model is improved, and the utilization rate of weight information is enhanced. Finally, Self-attention mechanism is used to improve the ability of model mining information. This paper selects the data sets of chndenticorp and CCF2012 to train the optimal value of the DenseNet feature mapping matrix. The optimal value is brought into the model contrast experiment. In the experiment, the accuracy rate, recall rate and F value of this method are all greater than 91%, which is the highest among the models. It effectively improves the accuracy of text sentiment analysis, and has high research and practical value.

Two problems arise when using distant supervision for relation extraction. First, in this method, an already existing knowledge base is heuristically aligned to texts, and the alignment results are treated as labeled data. However, the heuristic alignment can fail, resulting in wrong label problem. In addition, in previous approaches, statistical models have typically been applied to ad hoc features. The noise that originates from the feature extraction process can cause poor performance. In this paper, we propose a novel model dubbed the Piecewise Convolutional Neural Networks (PCNNs) with multi-instance learning to address these two problems. To solve the first problem, distant supervised relation extraction is treated as a multi-instance problem in which the uncertainty of instance labels is taken into account. To address the latter problem, we avoid feature engineering and instead adopt convolutional architecture with piecewise max pooling to automatically learn relevant features. Experiments show that our method is effective and outperforms several competitive baseline methods.

A Novel Automatic Image Annotation Method Based on Multi-Instance Learning

This paper presents a multimodal emotion recognition method based on dual-sequence Long Short-Term Memory (LSTM) network with attention mechanism. The ResNeXt50 network and coordinated attention mechanism are introduced to the video sequences to acquire the long-term dependence information in the location and space of video images. CNN with self-attention mechanism is utilized to catch the semantic features of audio sequences. In order to eliminate redundancy, a dual-sequence LSTM cross-modal network embedded with self-attention mechanism is applied for emotion feature fusion, and finally get the emotion output. The experimental results on RAVDESS dataset and eNTERFACE ’05 dataset prove that, compared with different baseline methods, the proposed algorithm is more attentive to the correlation between features and can significantly improve the recognition accuracy.

Multi-instance dictionary learning via multivariate performance measure optimization

Fine-grained classification of whole slide images (WSIs) is essential in precision oncology, enabling precise cancer diagnosis and personalized treatment strategies. The core of this task involves distinguishing subtle morphological variations within the same broad category of gigapixel-resolution images, which presents a significant challenge. While the multi-instance learning (MIL) paradigm alleviates the computational burden of WSIs, existing MIL methods often overlook hierarchical label correlations, treating fine-grained classification as a flat multi-class classification task. To overcome these limitations, we introduce a novel hierarchical multi-instance learning (HMIL) framework. By facilitating on the hierarchical alignment of inherent relationships between different hierarchy of labels at instance and bag level, our approach provides a more structured and informative learning process. Specifically, HMIL incorporates a class-wise attention mechanism that aligns hierarchical information at both the instance and bag levels. Furthermore, we introduce supervised contrastive learning to enhance the discriminative capability for fine-grained classification and a curriculum-based dynamic weighting module to adaptively balance the hierarchical feature during training. Extensive experiments on our large-scale cytology cervical cancer (CCC) dataset and two public histology datasets, BRACS and PANDA, demonstrate the state-of-the-art class-wise and overall performance of our HMIL framework. Our source code is available at https://github.com/ChengJin-git/HMIL.

Grain is a commodity related to the livelihood of the nation’s people, and the volatility of its futures price affects risk management, investment decisions, and policy making. Therefore, it is very necessary to establish an accurate and efficient futures price prediction model. Aiming at improving the accuracy and efficiency of the prediction model, so as to support reasonable decision making, this paper proposes a Bi-DSConvLSTM-Attention model for grain futures price prediction, which is based on the combination of a bidirectional long short-term memory neural network (BiLSTM), a depthwise separable convolutional long short-term memory neural network (DSConvLSTM), and an attention mechanism. Firstly, the mutual information is used to evaluate, sort, and select the features for dimension reduction. Secondly, the lightweight depthwise separable convolution (DSConv) is introduced to replace the standard convolution (SConv) in ConvLSTM without sacrificing its performance. Then, the self-attention mechanism is adopted to improve the accuracy. Finally, taking the wheat futures price prediction as an example, the model is trained and its performance is evaluated. Under the Bi-DSConvLSTM-Attention model, the experimental results of selecting the most relevant 1, 2, 3, 4, 5, 6, and 7 features as the inputs showed that the optimal number of features to be selected was 4. When the four best features were selected as the inputs, the RMSE, MAE, MAPE, and R2 of the prediction result of the Bi-DSConvLSTM-Attention model were 5.61, 3.63, 0.55, and 0.9984, respectively, which is a great improvement compared with the existing price-prediction models. Other experimental results demonstrated that the model also possesses a certain degree of generalization and is capable of obtaining positive returns.

One prominent feature of cognitive radio (CR) involves spectrum sensing (SS), which allows licensed primary users to remain unaffected by secondary users’ ability to discover and exploit unoccupied frequency bands. Spectrum sensing enhances the use of spectrum in CR devices, increasing their adaptability and efficiency in wireless communication systems. The rise of wireless equipment and the advent of IoT technologies compound this need for flexibility. Over time, the fixed allocation of frequencies has led to inefficiencies and underutilization as bandwidth needs increase. Deep learning and artificial intelligence have improved spectrum sensing by increasing detection probability of primary users’ presence under noisy environments, enabling cognitive radio systems to respond intelligently to fluctuations in RF environments. This article is concerned with deep learning techniques for spectrum sensing and modulation categorization with CBRT structure, which combines convolutional neural networks (CNNs), bidirectional recurrent neural networks (BRNNs), and transformer networks (TNs) to improve spectrum sensing. CNNs are responsible for performing spectrum feature extraction; BRNNs are used to capture temporal dependencies; and TNs are good at long range dependencies. Better performance for this model is aimed by integrating the three architectures described. In the proposed work, six digital modulation schemes were considered, for spectrum sensing. The sensing of spectrum in this model is performed using the RadioML2016.10B open-source dataset and performance metrics like the Jaccard Index (JI), Fowlkes’s Mallows Index, and F1 Score. Modulation classification has been performed using MIGOU-MOD open-source dataset. The proposed model exhibits good detection probability and low sensing error, unlike other methods at lower SNR.

This paper presents a new method for token-by-token sequence labeling that can leverage not only lexical information but also speech information without any alignments. Our motivation is to detect disfluencies such as fillers and word fragments robustly from spontaneous speech. Disfluency detection is often modeled as a token-by-token sequence labeling using a transcribed text via automatic speech recognition. However, utilizing the lexical information alone is not sufficient because the disfluencies cause changes to speech information. One problem is that the speech and the transcribed text need to be aligned when we handle speech and lexical information simultaneously. This prevents introducing speech information to the disfluency detection. To solve this problem, we propose a method for token-by-token sequence labeling, one that can simultaneously use lexical and speech information without requiring any alignments. To this end, we introduce attention mechanisms into a method for neural sequence labeling based on bi-directional long short-term memory recurrent neural network conditional random fields. The attention mechanisms enable us to find the term of disfluencies from speech automatically. Our experimental results show that the proposed method using acoustic and prosodic features improves the labeling accuracy compared with that using lexical features alone.