This paper addresses several key issues in the ArnetMiner system, which aims at extracting and mining academic social networks. Specifically, system focuses on: 1) Extracting researcher profiles automatically from Web; 2) Integrating publication data into network existing digital libraries; 3) Modeling entire network; 4) Providing search services for network. So far, 448,470 have been extracted using a unified tagging approach. We integrate publications online Web databases propose...

This paper explores the use of social annotations to improve websearch. Nowadays, many services, e.g. del.icio.us, have been developed for web users organize and share their favorite webpages on line by using annotations. We observe that can benefit search in two aspects: 1) are usually good summaries corresponding webpages; 2) count indicates popularity webpages. Two novel algorithms proposed incorporate above information into page ranking: SocialSimRank (SSR)calculates similarity between...

Retweeting is an important action (behavior) on Twitter, indicating the behavior that users re-post microblogs of their friends. While much work has been conducted for mining textual content generate or analyzing social network structure, few publications systematically study underlying mechanism retweeting behaviors. In this paper, we perform interesting analysis problem Twitter. We have found almost 25.5% tweets posted by are actually retweeted from friends' blog spaces. Our investigation...

As a social service in Web 2.0, folksonomy provides the users ability to save and organize their bookmarks online with "social annotations" or "tags". Social annotations are high quality descriptors of web pages' topics as well good indicators users' interests. We propose personalized search framework utilize for search. Specifically, three properties folksonomy, namely categorization, keyword, structure property, explored. In framework, rank page is decided not only by term matching between...

The boom of product review websites, blogs and forums on the web has attracted many research efforts opinion mining. Recently, there was a growing interest in finer-grained mining, which detects opinions different features as opposed to whole level. researches feature-level mining mainly rely identifying explicit relatedness between feature words reviews. However, sentiment two objects is usually complicated. For cases, are implied by detection such hidden association still big challenge...

This paper is concerned with the problem of mining social emotions from text. Recently, fast development web 2.0, more and documents are assigned by users emotion labels such as happiness, sadness, surprise. Such can provide a new aspect for document categorization, therefore help online to select related based on their emotional preferences. Useful it is, ratio manual still very tiny comparing huge amount web/enterprise documents. In this paper, we aim discover connections between affective...

In recent years, the number of freely available online reviews is increasing at a high speed. Aspect-based opinion mining technique has been employed to find out reviewers' opinions toward different product aspects. Such finer-grained valuable for potential customers make their purchase decisions. Product-feature extraction and categorization very important better aspect-oriented opinions. Since people usually use words describe same aspect in reviews, product-feature becomes more...

Sentiment analysis, which addresses the computational treatment of opinion, sentiment, and subjectivity in text, has received considerable attention recent years. In contrast to traditional coarse-grained sentiment analysis tasks, such as document-level classification, we are interested fine-grained aspect-based that aims identify aspects users comment on these aspects' polarities. Aspect-based relies heavily syntactic features. However, reviews this task focuses natural spontaneous, thus...

Mengting Hu, Shiwan Zhao, Li Zhang, Keke Cai, Zhong Su, Renhong Cheng, Xiaowei Shen. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and 9th International Joint (EMNLP-IJCNLP). 2019.

As an increasing number of users access information on the Web, there is a great opportunity to learn from server logs about users' probable actions in future. We present n-gram based model utilize path profiles very large data sets predict future requests. Since this prediction system, we cannot measure recall traditional sense. We, therefore, notion applicability give ability next document. Our simple extension existing point-based models for such predictions, but our results show when n...

This paper is concerned with the problem of browsing social annotations. Today, a lot services (e.g., Del.icio.us, Filckr) have been provided for helping users to manage and share their favorite URLs photos based on Due exponential increasing annotations, more users, however, are facing how effectively find desired resources from large annotation data. Existing methods such as tag cloud matching work well only small sets. Thus, an effective approach scale sets associated in great demand by...

We study a novel problem of social context summarization for Web documents. Traditional research has focused on extracting informative sentences from standard With the rapid growth online networks, abundant user generated content (e.g., comments) associated with documents is available. Which parts in document are users really caring about? How can we generate summaries by considering both informativeness and interests users? This paper explores such an approach modeling contexts into unified...

This paper focuses on analyzing and predicting not-answered questions in Community based Question Answering (CQA) services, such as Yahoo! Answers. In CQA users express their information needs by submitting natural language await answers from other human users. Comparing to receiving results web search engines using keyword queries, are likely get more specific answers, because answerers may catch the main point of question. However, one key problems this pattern is that sometimes no helps...

This paper is concerned with the problem of social affective text mining, which aims to discover connections between emotions and terms based on user-generated emotion labels. We propose a joint emotion-topic model by augmenting latent Dirichlet allocation an additional layer for modeling. It first generates set topics from emotions, followed generating each topic. Experimental results online news collection show that proposed can effectively identify meaningful emotion. Evaluation...

To recognize facial expression from candid, non-posed images, we propose a deep-learning based approach using convolutional neural networks (CNNs). In order to evaluate the performance in real-time candid recognition, have created image (CIFE) dataset, with seven types of more than 10,000 images gathered Web. As baselines, two feature-based approaches (LBP+SVM, SIFT+SVM) are tested on dataset. The structure our proposed CNN-based is described, and data augmentation technique provided...

Domain adaptation is an important problem in named entity recognition (NER). NER classifiers usually lose accuracy the domain transfer due to different data distribution between source and target domains. The major reason for performance degrading that each type often has lots of domain-specific term representations existing approaches need amount labeled tuning original model. However, it a labor-intensive time-consuming task build annotated training set every domain. We present method with...

Recently, deep convolutional neural networks (CNNs) have achieved great success in pathological image classification. However, due to the limited number of labeled images, there are still two challenges be addressed: (1) overfitting: performance a CNN model is undermined by overfitting its huge amounts parameters and insufficiency training data. (2) privacy leakage: trained using conventional method may involuntarily reveal private information patients dataset. The smaller dataset, worse...

Mengting Hu, Shiwan Zhao, Honglei Guo, Chao Xue, Hang Gao, Tiegang Renhong Cheng, Zhong Su. Proceedings of the 59th Annual Meeting Association for Computational Linguistics and 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Modern corporations operate in an extremely complex environment and strongly depend on all kinds of information resources across the enterprise. Unfortunately, with growth enterprise, its are not only heterogeneous but also distributed physically different systems databases. How to effectively exploit enterprise is becoming a critical hard problem. In recent years, metadata which detailed description data used efficiently web. The World Wide Web Consortium (W3C) recommends resource framework...

Opinion mining focuses on extracting customers' opinions from the reviews and predicting their sentiment orientation. Reviewers usually praise a product in some aspects bemoan it other aspects. With business globalization, is very important for enterprises to extract toward different find out cross-lingual/cross-culture difference opinions. Cross-lingual opinion challenging task as amounts of are written languages, not well structured. Since people use words describe same aspect reviews,...

Abstract Most current research on dynamic visual Simultaneous Localization and Mapping (SLAM) systems focuses scenes where static objects occupy most of the environment. However, in densely populated indoor environments, movement crowd can lead to loss feature information, thereby diminishing system’s robustness accuracy. This paper proposes a SLAM algorithm for dense environments based combination ORB-SLAM2 framework RGB-D cameras. Firstly, we introduced dedicated target detection network...

In the past few years, relevance feedback (RF) has been used as an effective solution for content-based image retrieval (CBIR). Although effective, RF-CBIR framework does not address issue of feature extraction dimension reduction and noise reduction. this paper, we propose a novel method extracting features class images represented by positive provided subjective RF. Principal Component Analysis (PCA) is to reduce both contained in original dimensionality spaces. The increases speed reduces...

The best known Scale-Invariant Feature Transform (SIFT) shows its superior performance in a variety of image processing tasks due to distinctiveness, invariance scale, rotation and local geometric distortion. Despite remarkable performance, SIFT is not invariant mirror images grayscale-inverted images.