Wenyi Huang

RefSeer: a citation recommendation system

Citations are important in academic dissemination. To help researchers check the completeness of citations while authoring a paper, we introduce a citation recommendation system called RefSeer. Researchers can use it to find related works to cited while authoring papers. It can also be used by reviewers to check the completeness of a papers references. RefSeer presents both topic based global recommendation and also citation-context based local recommendation. By evaluating the quality of recommendation, we show that such recommendation system can recommend citations with good precision and recall. We also show that our recommendation system is very efficient and scalable.

Towards building a scholarly big data platform: challenges, lessons and opportunities

We introduce a Big Data platform that provides various services for harvesting scholarly information and enabling efficient scholarly applications. The core architecture of the platform is built on a secured private cloud, crawls data using a scholarly focused crawler that leverages a dynamic scheduler, processes by utilizing a map reduce based crawl-extraction-ingestion (CEI) workflow, and is stored in distributed repositories and databases. Services such as scholarly data harvesting, information extraction, and user information and log data analytics are integrated into the platform and provided by an OAI and RESTful API. We also introduce a set of scholarly applications built on top of this platform including citation recommendation and collaborator discovery.

Measuring Prerequisite Relations Among Concepts

A prerequisite relation describes a basic relation among concepts in cognition, education and other areas. However, as a semantic relation, it has not been well studied in computational linguistics. We investigate the problem of measuring prerequisite relations among concepts and propose a simple link-based metric, namely reference distance (RefD), that effectively models the relation by measuring how differently two concepts refer to each other. Evaluations on two datasets that include seven domains show that our single metric based method outperforms existing supervised learning based methods.

MtNet: A Multi-Task Neural Network for Dynamic Malware Classification

In this paper, we propose a new multi-task, deep learning architecture for malware classification for the binary i.e. malware versus benign malware classification task. All models are trained with data extracted from dynamic analysis of malicious and benign files. For the first time, we see improvements using multiple layers in a deep neural network architecture for malware classification. The system is trained on 4.5 million files and tested on a holdout test set of 2 million files which is the largest study to date. To achieve a binary classification error rate of 0.358i¾?%, the objective functions for the binary classification task and malware family classification task are combined in the multi-task architecture. In addition, we propose a standard i.e. non multi-task malware family classification architecture which also achieves a malware family classification error rate of 2.94i¾?%.

Aggregating Local Context for Accurate Scene Text Detection

Scene text reading continues to be of interest for many reasons including applications for the visually impaired and automatic image indexing systems. Here we propose a novel end-to-end scene text detection algorithm. First, for identifying text regions we design a novel Convolutional Neural Network (CNN) architecture that aggregates local surrounding information for cascaded, fast and accurate detection. The local information serves as context and provides rich cues to distinguish text from background noises. In addition, we designed a novel grouping algorithm on top of detected character graph as well as a text line refinement step. Text line refinement consists of a text line extension module, together with a text line filtering and regression module. Jointly they produce accurate oriented text line bounding box. Experiments show that our method achieved state-of-the-art performance in several benchmark data sets: ICDAR 2003 (IC03), ICDAR 2013 (IC13) and Street View Text (SVT).

Detecting Arbitrary Oriented Text in the Wild with a Visual Attention Model

Text embedded in images provides important semantic information about a scene and its content. Detecting text in an unconstrained environment is a challenging task because of the many fonts, sizes, backgrounds, and alignments of the characters. We present a novel attention model for detecting arbitrary oriented and curved scene text. Inspired by the attention mechanisms in the human visual system, our model utilizes a spatial glimpse network to processes the attended area and deploys a recurrent neural network that aggregates the information over time to determine the attention movement. Combining this with an off-the-shelf region proposal method, the model achieves the state-of-the-art performance on the highly cited ICDAR2013 dataset, and the MSRA-TD500 dataset which contains arbitrary oriented text.

Crowd-sourcing web knowledge for metadata extraction

We explore a new metadata extraction framework without human annotators with the ground truth harvested from Web. A new training sample is selected based on not only the uncertainty and representativeness in the unlabeled pool, but also on its availability and credibility in Web knowledge bases. We construct a dataset of 4329 books with valid metadata and evaluate our approach using 5 Web book databases as oracles. Empirical results demonstrate its effectiveness and efficiency.

Smart library: identifying books on library shelves using supervised deep learning for scene text reading

Physical library collections are valuable and long standing resources for knowledge and learning. However, managing and finding books or other volumes on a large collection of bookshelves often leads to tedious manual work, especially for large collections where books or others might be missing or misplaced. Recently, deep neural-based models have been successful in detecting and recognizing text in images taken from natural scenes. Based on this, we investigate deep learning for facilitating book management. This task introduces further challenges including image distortion and varied lighting conditions. We present a library inventory building and retrieval system based on scene text reading. We specifically design our text recognition model using rich supervision to accelerate training and achieve state-of-the- art performance on several benchmark datasets. Our proposed system has the potential to greatly reduce the amount of manual labor required for managing book inventories.