This paper presents GRASTA (Grassmannian Robust Adaptive Subspace Tracking Algorithm), an efficient and robust online algorithm for tracking subspaces from highly incomplete information. The algorithm uses a robust $l^1$-norm cost function in order to estimate and track non-stationary subspaces when the streaming data vectors are corrupted with outliers. We apply GRASTA to the problems of robust matrix completion and real-time separation of background from foreground in video. In this second application, we show that GRASTA performs high-quality separation of moving objects from background at exceptional speeds: In one popular benchmark video example, GRASTA achieves a rate of 57 frames per second, even when run in MATLAB on a personal laptop.

The use of community detection algorithms is explored within the framework of cover song identification, i.e. the automatic detection of different audio renditions of the same underlying musical piece. Until now, this task has been posed as a typical query-by-example task, where one submits a query song and the system retrieves a list of possible matches ranked by their similarity to the query. In this work, we propose a new approach which uses song communities to provide more relevant answers to a given query. Starting from the output of a state-of-the-art system, songs are embedded in a complex weighted network whose links represent similarity (related musical content). Communities inside the network are then recognized as groups of covers and this information is used to enhance the results of the system. In particular, we show that this approach increases both the coherence and the accuracy of the system. Furthermore, we provide insight into the internal organization of individual cover song communities, showing that there is a tendency for the original song to be central within the community. We postulate that the methods and results presented here could be relevant to other query-by-example tasks.

Self-Organizing Maps are commonly used for unsupervised learning purposes. This paper is dedicated to the certain modification of SOM called SOMN (Self-Organizing Mixture Networks) used as a mechanism for representing grayscale digital images. Any grayscale digital image regarded as a distribution function can be approximated by the corresponding Gaussian mixture. In this paper, the use of SOMN is proposed in order to obtain such approximations for input grayscale images in unsupervised manner.

We consider the problem of speaker diarization, the problem of segmenting an audio recording of a meeting into temporal segments corresponding to individual speakers. The problem is rendered particularly difficult by the fact that we are not allowed to assume knowledge of the number of people participating in the meeting. To address this problem, we take a Bayesian nonparametric approach to speaker diarization that builds on the hierarchical Dirichlet process hidden Markov model (HDP-HMM) of Teh et al. [J. Amer. Statist. Assoc. 101 (2006) 1566--1581]. Although the basic HDP-HMM tends to over-segment the audio data---creating redundant states and rapidly switching among them---we describe an augmented HDP-HMM that provides effective control over the switching rate. We also show that this augmentation makes it possible to treat emission distributions nonparametrically. To scale the resulting architecture to realistic diarization problems, we develop a sampling algorithm that employs a truncated approximation of the Dirichlet process to jointly resample the full state sequence, greatly improving mixing rates. Working with a benchmark NIST data set, we show that our Bayesian nonparametric architecture yields state-of-the-art speaker diarization results.

Using these custom tools requires have been busily converting material from paper and microfilm training and a skilled workforce. We show in this paper that into digital domain. Newspapers, books, journals and some parts of that process can be distributed to a pool of even individual letters are finding themselves inside large unskilled volunteers with good results.

We evaluate the statistical dependency parser, Malt, on a new dataset of sentences taken from tweets. We use a version of Malt which is trained on gold standard phrase structure Wall Street Journal (WSJ) trees converted to Stanford labelled dependencies. We observe a drastic drop in performance moving from our in-domain WSJ test set to the new Twitter dataset, much of which has to do with the propagation of part-of-speech tagging errors. Retraining Malt on dependency trees produced by a state-of-the-art phrase structure parser, which has itself been self-trained on Twitter material, results in a significant improvement. We analyse this improvement by examining in detail the effect of the retraining on individual dependency types.

The new user problem (aka user cold start) is very common in online recommender systems. Active collaborative filtering (active CF) tries to solve this problem by intelligently soliciting user feedback in order to build an initial user profile with minimal costs. Existing methods only query the user for feedback on items, while users can have preferences over items as well as certain item attributes. In this paper, we extend active CF via user feedback on both items and attributes. For example, when making movie recommendations, the system can ask users for not only their favorite movies, but also attributes such as genres, actors, etc. We design a unified active CF framework for incorporating both item and attribute feedback based on the random walk model. We test the active CF algorithm on real-world movie recommendation data sets to demonstrate that appropriately querying for both item and feature feedback can significantly reduce the overall user effort measured in terms of number of queries. We show that we can achieve much better recommendation quality as compared to traditional active CF methods that support only item feedback.

With the growth of online information, many people are challenged in finding and reading the information most important for their interests. From 2008-2010 we built an experimental personalized news system where readers can subscribe to organized channels of information that are curated by experts. AI technology was employed to radically reduce the work load of curators and to efficiently present information to readers. The system has gone through three implementation cycles and processed over 16 million news stories from about 12,000 RSS feeds on over 8000 topics organized by 160 curators for over 600 registered readers. This paper describes the approach, engineering and AI technology of the system.

NewsFinder automates the steps involved in finding, selecting and publishing news stories that meet subjective judgments of relevance and interest to the Artificial Intelligence community. NewsFinder combines a broad search with AI-specific filters and incorporates a learning program whose judgment of interestingness of stories can be trained by feedback from readers. Since August, 2010, the program has been used to operate the AI in the News service that is part of the AAAI AITopics site.

My work aims to build a system to automatically predict the writing quality in scientific articles from two genres—academic publications and science journalism. Our goal is to employ these predictions for article recommendation systems and to provide feedback during writing.