Hongyun Bao

Detecting dynamic association among twitter topics

Over the last few years, Twitter is increasingly becoming an important source of up-to-date topics about what is happening in the world. In this paper, we propose a dynamic topic association detection model to discover relations between Twitter topics, by which users can gain insights into richer information about topics of interest. The proposed model utilizes a time constrained method to extract event-based spatio-temporal topic association, and constructs a dynamic temporal map to represent the obtained result. Experimental results show the improvement of the proposed model compared to static spatio-temporal method and co-occurrence method.

A neural network framework for relation extraction

Relation extraction is to identify the relationship of two given entities in the text. It is an important step in the task of knowledge extraction. Most conventional methods for the task of relation extraction focus on designing effective handcrafted features or learning a semantic representation of the whole sentence. Sentences with the same relationship always share the similar expressions. Besides, the semantic properties of given entities can also help to distinguish some confusing relations. Based on the above observations, we propose a neural network based framework for relation classification. It can simultaneously learn the relation patterns information and the semantic properties of given entities. In this framework, we explore two specific models: the CNN-based model and LSTM-based model. We conduct experiments on two public datasets: the SemEval-2010 Task8 dataset and the ACE05 dataset. The proposed method achieves the state-of-the-art result without using any external information. Additionally, the experimental results also show that our approach can represent the semantic relationship of the given entities effectively.

Joint Extraction of Entities and Relations Based on a Novel Tagging Scheme.

Joint extraction of entities and relations is an important task in information extraction. To tackle this problem, we firstly propose a novel tagging scheme that can convert the joint extraction task to a tagging problem. Then, based on our tagging scheme, we study different end-to-end models to extract entities and their relations directly, without identifying entities and relations separately. We conduct experiments on a public dataset produced by distant supervision method and the experimental results show that the tagging based methods are better than most of the existing pipelined and joint learning methods. Whats more, the end-to-end model proposed in this paper, achieves the best results on the public dataset.

How shall we catch people's concerns in micro-blogging?

In micro-blogging, people talk about their daily life and change minds freely, thus by mining peoples interest in micro-blogging, we will easily perceive the pulse of society. In this paper, we catch what people are caring about in their daily life by discovering meaningful communities based on probabilistic factor model (PFM). The proposed solution identifies peoples interest from their friendship and content information. Therefore, it reveals the behaviors of people in micro-blogging naturally. Experimental results verify the effectiveness of the proposed model and show peoples social life vividly.

Distant supervision for relation extraction with hierarchical selective attention

Distant supervised relation extraction is an important task in the field of natural language processing. There are two main shortcomings for most state-of-the-art methods. One is that they take all sentences of an entity pair as input, which would result in a large computational cost. But in fact, few of most relevant sentences are enough to recognize the relation of an entity pair. To tackle these problems, we propose a novel hierarchical selective attention network for relation extraction under distant supervision. Our model first selects most relevant sentences by taking coarse sentence-level attention on all sentences of an entity pair and then employs word-level attention to construct sentence representations and fine sentence-level attention to aggregate these sentence representations. Experimental results on a widely used dataset demonstrate that our method performs significantly better than most of existing methods.

Joint Extraction of Multiple Relations and Entities by Using a Hybrid Neural Network

This paper proposes a novel end-to-end neural model to jointly extract entities and relations in a sentence. Unlike most existing approaches, the proposed model uses a hybrid neural network to automatically learn sentence features and does not rely on any Natural Language Processing (NLP) tools, such as dependency parser. Our model is further capable of modeling multiple relations and their corresponding entity pairs simultaneously. Experiments on the CoNLL04 dataset demonstrate that our model using only word embeddings as input features achieves state-of-the-art performance.

A Bidirectional Hierarchical Skip-Gram model for text topic embedding

Taking advantage of the large scale corpus on the web to effectively and efficiently mine the topics within texts is an essential problem in the era of big data. We focus on the problem of learning text topic embedding in an unsupervised manner, which enjoys the properties of efficiency and scalability. Text topic embedding represents words and documents in a semantic topic space, in which the words and documents with similar topic will be embedded close to each other. When compared with conventional topic models, which implicitly capture the document-level word co-occurrence patterns, text topic embedding alleviates the data sparsity problem and captures the semantic relevance between different words and documents. To model text topic embedding, we propose a Bidirectional Hierarchical Skip-Gram model (BHSG) based on skip-gram model. BHSG includes two components: semantic generation module to learn semantic relevance between texts and topic enhance module to produce the text topic embedding based on text embedding learned in the former module. We evaluated our method on two kinds of topic-related tasks: text classification and information retrieval. The experimental results on four public datasets and one dataset we provide all demonstrate that our proposed method can achieve a better performance.

A Novel Hierarchical Convolutional Neural Network for Question Answering over Paragraphs

The question of classical Factoid Question Answering (FQA) task is always in the form of a single sentence. There also exists another kind of FQA task, whose question is a descriptive paragraph, such as quiz bowl question answering. Recently, some works try to automatically answer paragraph questions by applying machine learning methods. However, these methods neglect the correlation information between sentences in a paragraph and do not take full advantage of answer embedding information. In this paper, we propose a novel Hierarchical Convolutional Neural Network, called HCNN-E, to settle the task by considering ordinal information of sentences in paragraph and the information of answer embeddings. The experimental results on two public datasets demonstrate the effectiveness of proposed method, and the proposed method can achieve approximately 10% - 20% improvements, when comparing with the baselines.

Rigid or Flexible? A New Navigation Approach for Better Consumer Service Based on Knowledge Enhancement

With the rapid development of the Internet and E-commerce, online shopping sites are becoming a popular platform for products selling. Shopping sites such as amazon.com, dangdang.com provide consumers with a hierarchical navigation for selecting products easily from overwhelming amount of products. However, those man-made navigations are so general and professional that consumers still need to spend much time in filtering out their own undesired products personally. Shopping sites provide abundant textual product descriptions for most products, which describes the details of the product. In this paper, we propose a novel model to build a topic hierarchy from the detailed product descriptions, which can automatically model words into a tree structure by hierarchical Latent Dirichlet Allocation (hLDA), besides, our model can also augment words level allocations with the conceptual relation between words in WordNet automatically. Each node in the hierarchical tree contains some relevant keywords of product descriptions, thus clarifying the meaning of the concept in the node. Therefore, consumers can pick out their interested products by using the discovered descriptive and valuable navigation of products. The experimental results on amazon.com, one of the most popular shopping sites in America, demonstrate the efficiency and effectiveness of our proposed model.