Pedro J. Moreno

Topic segmentation with an aspect hidden Markov model

We present a novel probabilistic method for topic segmentation on unstructured text. One previous approach to this problem utilizes the hidden Markov model (HMM) method for probabilistically modeling sequence data [7]. The HMM treats a document as mutually independent sets of words generated by a latent topic variable in a time series. We extend this idea by embedding Hofmanns aspect model for text [5] into the segmenting HMM to form an aspect HMM (AHMM). In doing so, we provide an intuitive topical dependency between words and a cohesive segmentation model. We apply this method to segment unbroken streams of New York Times articles as well as noisy transcripts of radio programs on SpeechBot, an online audio archive indexed by an automatic speech recognition engine. We provide experimental comparisons which show that the AHMM outperforms the HMM for this task.

Semantic analysis of song lyrics

We explore the use of song lyrics for automatic indexing of music. Using lyrics mined from the Web, we apply a standard text processing technique to characterize their semantic content. We then determine artist similarity in this space. We found lyrics can be used to discover natural genre clusters. Experiments on a publicly available set of 399 artists showed that determining artist similarity using lyrics is better than random, but inferior to a state-of-the-art acoustic similarity technique. However the approaches made different errors, suggesting they could be profitably combined

Speechbot: an experimental speech-based search engine for multimedia content on the web

As the Web transforms from a text-only medium into a more multimedia-rich medium, the need arises to perform searches based on the multimedia content. In this paper, we present an audio and video search engine to tackle this problem. The engine uses speech recognition technology to index spoken audio and video files from the World Wide Web (WWW) when no transcriptions are available. If transcriptions (even imperfect ones) are available, we can also take advantage of them to improve the indexing process. Our engine indexes several thousand talk and news radio shows covering a wide range of topics and speaking styles from a selection of public Web sites with multimedia archives. Our Web site is similar in spirit to normal Web search sites; it contains an index, not the actual multimedia content. The audio from these shows suffers in acoustic quality due to bandwidth limitations, coding, compression, and poor acoustic conditions. Our word error rate (WER) results using appropriately trained acoustic models show remarkable resilience to the high compression, although many factors combine to increase the average WERs over standard broadcast news benchmarks. We show that, even if the transcription is inaccurate, we can still achieve good retrieval performance for typical user queries (77.5%).

From multimedia retrieval to knowledge management

Developments in information retrieval technologies can make multimedia data as pervasive and important as textual sources in knowledge management systems. The authors suggest ways in which speech-based multimedia information retrieval technologies can evolve into full-fledged knowledge management systems.

News Tuner: a simple interface for searching and browsing radio archives

We present in this paper a new Web-based application, called the News Tuner, for searching and browsing large radio archives. While popular search engines provide means for finding text and images, our approach combines semantic and acoustic search for efficient retrieval of audio documents. Semantic search allows the user to retrieve stories for a given concept, while acoustic search allows random access within stored audio files. Our experiments on over 1700 programs show that our method is effective at quickly retrieving stories that would be difficult to find otherwise. The News Tuner paradigm is intended primarily for news and talk radio programs, however it may be applied to browsing and searching any spoken word audio content

Reducing the cost of protein identifications from mass spectrometry databases

We present two techniques to improve the computational efficiency of protein discovery from mass spectrometry databases: noise filtering and hierarchical searching. Our approaches are orthogonal to existing algorithms and are based on the observation that typical mass spectrometry data contains a large amount of noise that can lead to wasteful computation. Our first improvement uses standard machine learning techniques with novel feature vectors derived from the mass spectra to identify and filter the noisy spectra. We demonstrate this approach results in computational gains of around 38% with less than 10% loss of peptides. Additionally we present a hierarchical searching scheme in which most samples are matched against a small database at low computational cost, leaving only a small number of samples to be searched against larger databases. Combining this scheme with the machine learning filters leads to a further performance improvement of 3%.

Protein seer: a Web server for protein homology detection

We present and evaluate a publicly available Web server which classifies protein sequences into SCOP 1.63 PDB95 structural superfamilies. The Website returns ranked lists of likely superfamilies and hence implicit structural predictions according to three computational techniques: BLAST, HMMER and a discriminative classifier SVM-BLOCKS. It is the first Website to provide predictions using SVM-BLOCKS. In addition to the ranked lists, the Website displays alignment information and a Web services interface is also available for computationally intensive use. We conduct a large-scale evaluation which mimics the predictions returned by the Website. The study indicates that the site provides valid predictions and that SVM-BLOCKS approach can outperform BLAST and HMMER when sufficient examples are available to learn the SVM classifiers.