Abstract-We study the problem of predicting home locations of Twitter users using contents of their tweet messages. Using three probability models for locations, we compare both the Gaussian Mixture Model (GMM) and the Maximum Likelihood Estimation (MLE). In addition, we propose two novel unsupervised methods based on the notions of Non-Localness and Geometric-Localness to prune noisy data from tweet messages. In the experiments, our unsupervised approach improves the baselines significantly and shows comparable results with the supervised state-of-the-art method. For 5,113 Twitter users in the test set, on average, our approach with only 250 selected local words or less is able to predict their home locations (within 100 miles) with the accuracy of 0.499, or has 509.3 miles of average error distance at best.

Abstract. Finding near-duplicate images is a task often found in Multimedia Information Retrieval (MIR). Toward this effort, we propose a novel idea by bridging two seemingly unrelated fields -MIR and Biology. That is, we propose to use the popular gene sequence alignment algorithm in Biology, i.e., BLAST, in detecting near-duplicate images. Under the new idea, we study how various image features and gene sequence generation methods (using gene alphabets such as A, C, G, and T in DNA sequences) affect the accuracy and performance of detecting near-duplicate images. Our proposal, termed as BLASTed Image Linkage (BASIL), is empirically validated using various real data sets. This work can be viewed as the "first" step toward bridging MIR and Biology fields in the well-studied near-duplicate image detection problem.

No abstract

Abstract-With the advance of Web technologies, the number of "academic" videos available on the Web (e.g., online lectures, web seminars, conference presentations, or tutorial videos) has increased explosively. A fundamental task of managing such videos is to classify them into relevant subjects. For this task, most of current content providers rely on keywords to perform the classification, while active techniques for automatic video classification focus on utilizing multi-modal features. However, in our settings, we argue that both approaches are not sufficient to solve the problem effectively. Keywords based method is very limited in terms of accuracy, while features based one lacks semantics to represent academic subjects. Toward this problem, in this paper, we propose to transform the video subject classification problem into the text categorization problem by exploiting the extracted transcripts of videos. Using both real and synthesized data, (1) we extensively study the validity of the proposed idea, (2) we analyze the performance of different text categorization methods, and (3) we study the impact of various factors of transcripts such as quality and length towards academic video classification problem.