Qiang Qu

A Topic Drift Model for authorship attribution

Authorship attribution is an active research direction due to its legal and financial importance. Its goal is to identify the authorship from the anonymous texts. In this paper, we propose a Topic Drift Model (TDM), which can monitor the dynamicity of authors writing styles and learn authors interests simultaneously. Unlike previous authorship attribution approaches, our model is sensitive to the temporal information and the ordering of words. Thus it can extract more information from texts. The experimental results show that our model achieves better results than other models in terms of accuracy. We also demonstrate the potential of our model to address the authorship verification problem.

Feature-enhanced attention network for target-dependent sentiment classification

Abstract In this paper, we propose a Feature-enhanced Attention Network to improve the performance of target-dependent Sentiment classification (FANS). Specifically, we first learn the feature-enhanced word representations by leveraging the unigram features, part of speech features and word position features. Second, we develop an multi-view co-attention network to learn a better multi-view sentiment-aware and target-specific sentence representation via interactively modeling the context words, target words and sentiment words. We conduct experiments to verify the effectiveness of our model on two real-world datasets in both English and Chinese. The experimental results demonstrate that FANS has robust superiority over competitors and sets state-of-the-art.

Cross-domain Aspect/Sentiment-aware Abstractive Review Summarization

This study takes the lead to study the aspect/sentiment-aware abstractive review summarization in domain adaptation scenario. The proposed model CASAS (neural attentive model for Cross-domain Aspect/Sentiment-aware Abstractive review Summarization) leverages domain classification task, working on datasets of both source and target domains, to recognize the domain information of texts and transfer knowledge from source domains to target domains. The extensive experiments on Amazon reviews demonstrate that CASAS outperforms the compared methods in both out-of-domain and in-domain setups.

GPSP: Graph Partition and Space Projection based Approach for Heterogeneous Network Embedding

In this paper, we propose GPSP, a novel Graph Partition and Space Projection based approach, to learn the representation of a heterogeneous network that consists of multiple types of nodes and links. Concretely, we first partition the heterogeneous network into homogeneous and bipartite subnetworks. Then, the projective relations hidden in bipartite subnetworks are extracted by learning the projective embedding vectors. Finally, we concatenate the projective vectors from bipartite subnetworks with the ones learned from homogeneous subnetworks to form the final representation of the heterogeneous network. Extensive experiments are conducted on a real-life dataset. The results demonstrate that GPSP outperforms the state-of-the-art baselines in two key network mining tasks: node classification and clustering.

MARES: multitask learning algorithm for Web-scale real-time event summarization

Automatic real-time summarization of massive document streams on the Web has become an important tool for quickly transforming theoverwhelming documents into a novel, comprehensive and concise overview of an event for users. Significant progresses have been made in static text summarization. However, most previous work does not consider the temporal features of the document streams which are valuable in real-time event summarization. In this paper, we propose a novel M ultitask learning A lgorithm for Web-scale R eal-time E vent S ummarization (MARES), which leverages the benefits of supervised deep neural networks as well as a reinforcement learning algorithm to strengthen the representation learning of documents. Specifically, MARES consists two key components: (i) A relevance prediction classifier, in which a hierarchical LSTM model is used to learn the representations of queries and documents; (ii) A document filtering model learns to maximize the long-term rewards with reinforcement learning algorithm, working on a shared document encoding layer with the relevance prediction component. To verify the effectiveness of the proposed model, extensive experiments are conducted on two real-life document stream datasets: TREC Real-Time Summarization Track data and TREC Temporal Summarization Track data. The experimental results demonstrate that our model can achieve significantly better results than the state-of-the-art baseline methods.

Multi-task Learning for Author Profiling with Hierarchical Features

Author profiling is an important but challenging task. In this paper, we propose a novel Multi-Task learning framework for Author Profiling (MTAP), in which a document modeling module is shared across three different author profiling tasks (i.e., age, gender and job classification tasks). To further boost author profiling, we integrate hierarchical features learned by different models. Concretely, we employ CNN, LSTM and topic model to learn the character-level, word-level and topic-level features, respectively. MTAP thus leverages the benefits of supervised deep neural neural networks as well as an unsupervised probabilistic generative model to enhance the document representation learning. Experimental results on a real-life blog dataset show that MTAP has robust superiority over competitors and sets state-of-the-art for all the three author profiling tasks

Ramp loss one-class support vector machine; A robust and effective approach to anomaly detection problems

Abstract Anomaly detection defines as a problem of finding those data samples, which do not follow the patterns of the majority of data points. Among the variety of methods and algorithms proposed to deal with this problem, boundary based methods include One-class support vector machine (OC-SVM) is considered as an effective and outstanding one. Nevertheless, extremely sensitivity to the presence of outliers and noises in the training set is considered as an important drawback of this group of classifiers. In this paper, we address this problem by developing a robust and sparse methodology for anomaly detection by introducing Ramp loss function to the original One-class SVM, called “Ramp-OCSVM”. The main objective of this research is to taking the advantages of non-convexity properties of the Ramp loss function to make robust and sparse semi-supervised algorithm. Furthermore, the Concave–Convex Procedure (CCCP) is utilized to solve the obtained model that is a non-differentiable non-convex optimization problem. We do comprehensive experiments and parameters sensitivity analysis on two artificial data sets and some chosen data sets from UCI repository, to show the superiority of our model in terms of detection power and sparsity. Moreover, some evaluations are done with NSL-KDD and UNSW-NB15 data sets as well-known and recently published intrusion detection data sets, respectively. The obtained results reveal the outperforming of our model in terms of robustness to outliers and superiority in the detection of anomalies.

Personalized response generation by Dual-learning based domain adaptation

Open-domain conversation is one of the most challenging artificial intelligence problems, which involves language understanding, reasoning, and the utilization of common sense knowledge. The goal of this paper is to further improve the response generation, using personalization criteria. We propose a novel method called PRGDDA (Personalized Response Generation by Dual-learning based Domain Adaptation) which is a personalized response generation model based on theories of domain adaptation and dual learning. During the training procedure, PRGDDA first learns the human responding style from large general data (without user-specific information), and then fine-tunes the model on a small size of personalized data to generate personalized conversations with a dual learning mechanism. We conduct experiments to verify the effectiveness of the proposed model on two real-world datasets in both English and Chinese. Experimental results show that our model can generate better personalized responses for different users.

Task-oriented keyphrase extraction from social media

Keyphrase extraction from social media is a crucial and challenging task. Previous studies usually focus on extracting keyphrases that provide the summary of a corpus. However, they do not take users’ specific needs into consideration. In this paper, we propose a novel three-stage model to learn a keyphrase set that represents or related to a particular topic. Firstly, a phrase mining algorithm is applied to segment the documents into human-interpretable phrases. Secondly, we propose a weakly supervised model to extract candidate keyphrases, which uses a few pre-specific seed keyphrases to guide the model. The model consequently makes the extracted keyphrases more specific and related to the seed keyphrases (which reflect the user’s needs). Finally, to further identify the implicitly related phrases, the PMI-IR algorithm is employed to obtain the synonyms of the extracted candidate keyphrases. We conducted experiments on two publicly available datasets from news and Twitter. The experimental results demonstrate that our approach outperforms the state-of-the-art baselines and has the potential to extract high-quality task-oriented keyphrases.