Victor Soto

Characterizing Urban Landscapes Using Geolocated Tweets

The pervasiveness of cell phones and mobile social media applications is generating vast amounts of geolocalized user-generated content. Since the addition of geotagging information, Twitter has become a valuable source for the study of human dynamics. Its analysis is shedding new light not only on understanding human behavior but also on modeling the way people live and interact in their urban environments. In this paper, we evaluate the use of geolocated tweets as a complementary source of information for urban planning applications. Our contributions are focussed in two urban planing areas: (1) a technique to automatically determine land uses in a specific urban area based on tweeting patterns, and (2) a technique to automatically identify urban points of interest as places with high activity of tweets. We apply our techniques in Manhattan (NYC) using 49 days of geolocated tweets and validate them using land use and landmark information provided by various NYC departments. Our results indicate that geolocated tweets are a powerful and dynamic data source to characterize urban environments.

Automated land use identification using cell-phone records

Pervasive large-scale infrastructures generate datasets that contain human behavior info rmation. In this context, cell phones and cell phone networks, due to its pervasiveness, can be considered sensors of human behavior and one of the main elements that define our digital footprint. In this paper we present a technique for the automatic identification and classification of land uses from the information generated by a cell-phone network infrastructure. Our approach first computes the aggregated calling patterns of the antennas of the network and, after that, finds the optimum cluster distribution to automatically identify how citizens use the different geographic regions within a city. We present and validate our results using cell phone records collected for the city of Madrid.

Prediction of socioeconomic levels using cell phone records

The socioeconomic status of a population or an individual provides an understanding of its access to housing, education, health or basic services like water and electricity. In itself, it is also an indirect indicator of the purchasing power and as such a key element when personalizing the interaction with a customer, especially for marketing campaigns or offers of new products. In this paper we study if the information derived from the aggregated use of cell phone records can be used to identify the socioeconomic levels of a population. We present predictive models constructed with SVMs and Random Forests that use the aggregated behavioral variables of the communication antennas to predict socioeconomic levels. Our results show correct prediction rates of over 80% for an urban population of around 500,000 citizens.

A double pruning algorithm for classification ensembles

This article introduces a double pruning algorithm that can be used to reduce the storage requirements, speed-up the classification process and improve the performance of parallel ensembles. A key element in the design of the algorithm is the estimation of the class label that the ensemble assigns to a given test instance by polling only a fraction of its classifiers. Instead of applying this form of dynamical (instance-based) pruning to the original ensemble, we propose to apply it to a subset of classifiers selected using standard ensemble pruning techniques. The pruned subensemble is built by first modifying the order in which classifiers are aggregated in the ensemble and then selecting the first classifiers in the ordered sequence. Experiments in benchmark problems illustrate the improvements that can be obtained with this technique. Specifically, using a bagging ensemble of 101 CART trees as a starting point, only the 21 trees of the pruned ordered ensemble need to be stored in memory. Depending on the classification task, on average, only 5 to 12 of these 21 classifiers are queried to compute the predictions. The generalization performance achieved by this double pruning algorithm is similar to pruned ordered bagging and significantly better than standard bagging.

Rescoring Confusion Networks for Keyword Search

We introduce a two-stage cascaded scheme to rescore Confusion Networks (CNs) for Keyword Search in the context of Low-Resource Languages. In the first stage we rescore the CN to improve the error rate of the 1-best hypothesis using a large number of lexical, phonetic, false alarms and structural features. Using a rank learning Support Vector Machine classifier, we obtain WER gains between 0.54% and 2.84% on Cantonese, Tagalog, Turkish, Pashto and Vietnamese. In the second stage we generate keyword hits from the rescored CN and use logistic regression to detect true hits and false alarms. We compare these to hits generated from the unrescored CN and obtain gains between 0.45% and 0.9% on the MTWV metric by using the mentioned features and including acoustic and prosodic features on Tagalog, Turkish and Pashto.

A double pruning scheme for boosting ensembles.

Ensemble learning consists of generating a collection of classifiers whose predictions are then combined to yield a single unified decision. Ensembles of complementary classifiers provide accurate and robust predictions, which are often better than the predictions of the individual classifiers in the ensemble. Nevertheless, ensembles also have some drawbacks: typically, all classifiers are queried to compute the final ensemble prediction. Therefore, all the classifiers need to be accessible to address potential queries. This entails larger storage requirements and slower predictions than a single classifier. Ensemble pruning techniques are useful to alleviate these drawbacks. Static pruning techniques reduce the ensemble size by selecting a sub-ensemble of classifiers from the original ensemble. In dynamic pruning, the querying process is halted when the partial ensemble prediction is sufficient to reach a stable final decision with a reasonable amount of confidence. In this paper, we present the results of a comprehensive analysis of static and dynamic pruning techniques applied to Adaboost ensembles. These ensemble pruning techniques are evaluated on a wide range of classification problems. From this analysis, one concludes that the combination of static and dynamic pruning techniques provides a notable reduction in the memory requirements and an improvement in the classification time without a significant loss of prediction accuracy.

Cross-language phrase boundary detection

We describe models of prosodic phrasing trained on multiple languages to identify boundaries in an unseen language. Our goal is to create models from High Resource languages, in which hand-annotated prosodic phrase boundaries are available, to use in identifying boundaries in a Low Resource language, with little or no training material. We train models on American English, Italian, Mandarin, and German and test on each of these languages. We find that, while pause is the most important feature for phrase boundary prediction in all languages examined, the role of pause in boundary identification varies by annotator and the relative importance of other features varies significantly by language. We also find that different acoustic correlates of prosodic boundaries characterize different languages. In some, the relative importance of features is silence > pitch > intensity > duration, while for other languages intensity is more important than pitch. These differences do not appear to be attributable to language family, since, e.g. English and German display different patterns.

Selection and combination of hypotheses for dialectal speech recognition

While research has often shown that building dialect-specific Automatic Speech Recognizers is the optimal approach to dealing with dialectal variations of the same language, we have observed that dialect-specific recognizers do not always output the best recognitions. Often enough, another dialectal recognizer outputs a better recognition than the dialect-specific one. In this paper, we present two methods to select and combine the best decoded hypothesis from a pool of dialectal recognizers. We follow a Machine Learning approach and extract features from the Speech Recognition output along with Word Embeddings and use Shallow Neural Networks for classification. Our experiments using Dictation and Voice Search data from the main four Arabic dialects show good WER improvements for the hypothesis selection scheme, reducing the WER by 2.1 to 12.1% depending on the test set, and promising results for the hypotheses combination scheme.

Consensus clustering for urban land use analysis using cell phone network data

Pervasive large-scale infrastructures have the ability to capture individual digital footprints and, as a result, provide a new vision on human dynamics. In this context, cell phones and cell phone networks, due to its ubiquity, can be considered one of the main sensors of human behaviour. The information collected by these networks can be used to understand the dynamics of urban environments with a detail not available up to now. One of the areas that can benefit from this information is urban planning. In this paper, we present a technique for the automatic identification of land uses from the information gathered by a cell phone network. Given the inherent diversity of human activities, we use consensus clustering to identify land uses, characterising only those geographical areas with well-defined behaviours. We present and validate our results using cell phone records and official land use data collected for Madrid.

Part of Speech Tagging for Code Switched Data.

We address the problem of Part of Speech tagging (POS) in the context of linguistic code switching (CS). CS is the phenomenon where a speaker switches between two languages or variants of the same language within or across utterances, known as intra-sentential or inter-sentential CS, respectively. Processing CS data is especially challenging in intra-sentential data given state of the art monolingual NLP technology since such technology is geared toward the processing of one language at a time. In this paper we explore multiple strategies of applying state of the art POS taggers to CS data. We investigate the landscape in two CS language pairs, Spanish-English and Modern Standard Arabic-Arabic dialects. We compare the use of two POS taggers vs. a unified tagger trained on CS data. Our results show that applying a machine learning framework using two state of the art POS taggers achieves better performance compared to all other approaches that we investigate.