Towards Accurate and Consistent Evaluation: A Dataset for Distantly-Supervised Relation Extraction

In the last decades, named entity recognition has been extensively studied with various supervised learning approaches depend on massive labeled data. In this paper, we focus on person name recognition in judgment documents. Owing to the lack of human-annotated data, we propose a joint learning approach, namely Aux-LSTM, to use a large scale of auto-annotated data to help human-annotated data (in a small size) for person name recognition. Specifically, our approach first develops an auxiliary Long Short-Term Memory (LSTM) representation by training the auto-annotated data and then leverages the auxiliary LSTM representation to boost the performance of classifier trained on the human-annotated data. Empirical studies demonstrate the effectiveness of our proposed approach to person name recognition in judgment documents with both human-annotated and auto-annotated data.

Distant supervision for relation extraction with hierarchical selective attention

Deep convolutional neural networks have achieved great success in recent years. They have been widely used in various applications such as optical and SAR image scene classification, object detection and recognition, semantic segmentation, and change detection. However, deep neural networks rely on large-scale high-quality training data, and can only guarantee good performance when the training and test data are independently sampled from the same distribution. Deep convolutional neural networks are found to be vulnerable to subtle adversarial perturbations. This adversarial vulnerability prevents the deployment of deep neural networks in security-sensitive applications such as medical, surveillance, autonomous driving and military scenarios. This paper first presents a holistic view of security issues for deep convolutional neural network-based image recognition systems. The entire information processing chain is analyzed regarding safety and security risks. In particular, poisoning attacks and evasion attacks on deep convolutional neural networks are analyzed in detail. The root causes of adversarial vulnerabilities of deep recognition models are also discussed. Then, we give a formal definition of adversarial robustness and present a comprehensive review of adversarial attacks, adversarial defense, and adversarial robustness evaluation. Rather than listing existing research, we focus on the threat models for the adversarial attack and defense arms race. We perform a detailed analysis of several representative adversarial attacks on SAR image recognition models and provide an example of adversarial robustness evaluation. Finally, several open questions are discussed regarding recent research progress from our workgroup. This paper can be further used as a reference to develop more robust deep neural network-based image recognition models in dynamic adversarial scenarios.

Recently, there is an effort to extend fine-grained entity typing by using a richer and ultra-fine set of types, and labeling noun phrases including pronouns and nominal nouns instead of just named entity mentions. A key challenge for this ultra-fine entity typing task is that human annotated data are extremely scarce, and the annotation ability of existing distant or weak supervision approaches is very limited. To remedy this problem, in this paper, we propose to obtain training data for ultra-fine entity typing by using a BERT Masked Language Model (MLM). Given a mention in a sentence, our approach constructs an input for the BERT MLM so that it predicts context dependent hypernyms of the mention, which can be used as type labels. Experimental results demonstrate that, with the help of these automatically generated labels, the performance of an ultra-fine entity typing model can be improved substantially. We also show that our approach can be applied to improve traditional fine-grained entity typing after performing simple type mapping.

Attention mechanisms are often used in deep neural networks for distantly supervised relation extraction (DS-RE) to distinguish valid from noisy instances. However, traditional 1- D vector attention models are insufficient for the learning of different contexts in the selection of valid instances to predict the relationship for an entity pair. To alleviate this issue, we propose a novel multi-level structured (2-D matrix) self-attention mechanism for DS-RE in a multi-instance learning (MIL) framework using bidirectional recurrent neural networks. In the proposed method, a structured word-level self-attention mechanism learns a 2-D matrix where each row vector represents a weight distribution for different aspects of an instance regarding two entities. Targeting the MIL issue, the structured sentence-level attention learns a 2-D matrix where each row vector represents a weight distribution on selection of different valid instances. Experiments conducted on two publicly available DS-RE datasets show that the proposed framework with a multi-level structured self-attention mechanism significantly outperform state-of-the-art baselines in terms of PR curves, P@N and F1 measures.

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.

Passive sensing applications are limited by their inability to determine who is using a device, a critical concern in child mobile device use research, where devices are often shared between siblings or between a child and their parent. Our previous work leveraged behavioral biometrics to identify a target child user; however, it is unknown what type of training data is necessary for optimal model performance. This study evaluated model performance across different characteristics of training data. Thirty-six children (11.3 ± 0.9 years, 56% female) self-selected a video or a game on iPads for 10 min while laying and for another 5 min while sitting. The SensorLog application captured iPad accelerometer and gyroscope data while the child interacted with the device. Machine learning algorithms including Neural Network (NN), Random Forest (RF), k-Nearest Neighbors (k-NN), and SwipeFormer were applied to determine the most important aspects of training data to optimize model performance. The aspects of training data evaluated included (1) varying the length (i.e., seconds of training data), (2) varying the user position (i.e., sitting, laying), and (3) varying the time proximity between training and testing data. F1 score was used to evaluate model performance. The SwipeFormer F1 scores were lowest when the training data was further from the test data (0 when training data was 11 min away from test data) and highest when training data was close to test data (0.91 when training data was the minute preceding test data). The SwipeFormer F1 scores were highest when predicting the user laying from laying (0.97) and sitting from sitting (0.94), and lowest when predicting the user sitting from laying (0) and laying from sitting (0). The length of training data had little impact on performance, with a SwipeFormer F1 score of 0.91 when training on one minute of data and a SwipeFormer F1 score of 0.94 when training on twelve minutes of data. Because researchers would likely be predicting users at different timepoints than their training data, research should focus on improving model performance for identifying users independent of time proximity for training and test data.

Supervised machine learning methods have been widely used in relation extraction to find the relation between two named entities in a sentence. However, the disadvantages of supervised machine learning methods are that constructing the training data set is costly and time-consuming, and the machine learning system is ultimately dependent on the specific domain of the training data. To overcome these disadvantages, we propose a two-step relation extraction model with distant supervision. The two-step model consists of a one-class model and a multi-class model. The one-class model selects positive sentences from input sentences and the multi-class model classifies the positive sentences into specific classes. In the experiments, the proposed model showed good F1-measures (62.9 % in the auto-labeled test data, 63.8 % in the gold-labeled test data), although it does not use any human-labeled training data.

Semantic piecewise convolutional neural network with adaptive negative training for distantly supervised relation extraction

Nitrate is a common pollutant in surface waters which water companies must monitor for regulatory and safety reasons. The presence of nitrate in deionised water is detected and concentration estimated from microwave spectroscopy measurements in the range 9kHz-6GHz. Experimental results were obtained for 19 solutions (18 salt solutions in deionised water and 1 deionised water), each measured 10 times with 4001 points (total N=190). The resulting data was randomly assigned into equal parts training and test data (N=95 each). Both regression (for the estimation of nitrate concentration) and classification (for detecting the presence of nitrate) methods were considered, with a rigorous feature selection procedure used to identify two frequencies for each of the classification and regression problems. For detection classification models were applied with nitrate levels binned using 30mg/l as the threshold. A logistic regression model achieved AUROC of 0.9875 on test data and a multi-layer perceptron achieved AUROC of 0.9871. In each case the positive predictive value of the model could be optimised at 100% with sensitivity of 90% and specificity of 100%. For the concentration estimates, a linear regression model was able to explain 42% of the variance in the training data and 45% of the variance in the test data and an MLP model delivered similar performance, explaining 43% of variance in the training data and 47% of variance in the test data. A sensor based on this model would be appropriate for detecting the presence of nitrate above a given threshold but poor at estimating concentration.

Face Anti-Spoofing (FAS) is essential to secure face recognition systems and has been extensively studied in recent years. Although deep neural networks (DNNs) for the FAS task have achieved promising results in intra-dataset experiments with similar distributions of training and testing data, the DNNs' generalization ability is limited under the cross-domain scenarios with different distributions of training and testing data. To improve the generalization ability, recent hybrid methods have been explored to extract task-aware handcrafted features (e.g., Local Binary Pattern) as discriminative information for the input of DNNs. However, the handcrafted feature extraction relies on experts' domain knowledge, and how to choose appropriate handcrafted features is underexplored. To this end, we propose a learnable network to extract Meta Pattern (MP) in our learning-to-learn framework. By replacing handcrafted features with the MP, the discriminative information from MP is capable of learning a more generalized model. Moreover, we devise a two-stream network to hierarchically fuse the input RGB image and the extracted MP by using our proposed Hierarchical Fusion Module (HFM). We conduct comprehensive experiments and show that our MP outperforms the compared handcrafted features. Also, our proposed method with HFM and the MP can achieve state-of-the-art performance on two different domain generalization evaluation benchmarks.

Distant supervision is a hotspot in relation extraction research. Instead of relying on annotated text, distant supervision hires a knowledge base as supervision. For each pair of entities that appears in some knowledge base's relation, this approach find all sentences containing those entities in a large unlabeled corpus and extract textual features to train a relation classifier. The automatic labeling provides a large amount of data, but the data have serious problem. Most features appear only few times in training data, and such insufficient data make these features very susceptible to noise, which will lead to a flawed classifier. In this paper, we propose a method to improve few occurrence features' performance in distant supervision relation extraction. We present a novel model to calculating the similarity between a feature and an entity pair, and then adjust the entity pair' features by their similarity. The experiment shows our method boosted the performance of relation extraction.

SummaryAccurate prediction of the solubility of chemical substances in solvents remains a challenge. The sparsity of high-quality solubility data is recognized as the biggest hurdle in the development of robust data-driven methods for practical use. Nonetheless, the effects of the quality and quantity of data on aqueous solubility predictions have not yet been scrutinized. In this study, the roles of the size and the quality of data sets on the performances of the solubility prediction models are unraveled, and the concepts of actual and observed performances are introduced. In an effort to curtail the gap between actual and observed performances, a quality-oriented data selection method, which evaluates the quality of data and extracts the most accurate part of it through statistical validation, is designed. Applying this method on the largest publicly available solubility database and using a consensus machine learning approach, a top-performing solubility prediction model is achieved.

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.

Distant supervised relation extraction has been widely used to find novel relational facts from text. However, distant supervision inevitably accompanies with the wrong labelling problem, and these noisy data will substantially hurt the performance of relation extraction. To alleviate this issue, we propose a sentence-level attention-based model for relation extraction. In this model, we employ convolutional neural networks to embed the semantics of sentences. Afterwards, we build sentence-level attention over multiple instances, which is expected to dynamically reduce the weights of those noisy instances. Experimental results on real-world datasets show that, our model can make full use of all informative sentences and effectively reduce the influence of wrong labelled instances. Our model achieves significant and consistent improvements on relation extraction as compared with baselines.