Learn More
Understanding how words change their meanings over time is key to models of language and cultural evolution, but historical data on meaning is scarce, making theories hard to develop and test. Word embeddings show promise as a di-achronic tool, but have not been carefully evaluated. We develop a robust methodology for quantifying semantic change by(More)
Probabilistic latent-variable models are a powerful tool for modelling structured data. However, traditional expectation-maximization methods of learning such models are both computationally expensive and prone to local-minima. In contrast to these traditional methods, recently developed learning algorithms based upon the method of moments are both(More)
Predictive state representations (PSRs) offer an expressive framework for modelling partially observable systems. By compactly representing systems as functions of observable quantities, the PSR learning approach avoids using local-minima prone expectation-maximization and instead employs a globally optimal moment-based algorithm. Moreover, since PSRs do(More)
The most polymorphic gene family in P. falciparum is the ∼60 var genes distributed across parasite chromosomes, both in the subtelomeres and in internal regions. They encode hypervariable surface proteins known as P. falciparum erythrocyte membrane protein 1 (PfEMP1) that are critical for pathogenesis and immune evasion in Plasmodium falciparum. How var(More)
Efficiently learning accurate models of dy-namical systems is of central importance for developing rational agents that can succeed in a wide range of challenging domains. The difficulty of this learning problem is particularly acute in settings with large observation spaces and partial observability. We present a new algorithm, called Compressed Predictive(More)
Pathogen genome sequencing directly from clinical samples is quickly gaining importance in genetic and medical research studies. However, low DNA yield from blood-borne pathogens is often a limiting factor. The problem worsens in extremely base-biased genomes such as the AT-rich Plasmodium falciparum. We present a strategy for whole-genome amplification(More)
OBJECTIVE To assess the effect of a comprehensive tobacco control programme initiated in Massachusetts in 1993, and to compare the 1990 to 1999 trend in smoking prevalence to that in 41 states without tobacco control programmes, controlling for demographic shifts over time. DESIGN Data from the Behavioral Risk Factor Surveillance System for the years 1990(More)
Smoking behavior has been shown to be influenced by individuals' perceptions of social norms about smoking. This study examines whether local regulations regarding clean indoor air and youth access to tobacco are associated with residents' subsequent perceptions of smoking norms. Data came from Massachusetts surveys of adults and youths and from records of(More)
A word's sentiment depends on the domain in which it is used. Computational social science research thus requires sentiment lexicons that are specific to the domains being studied. We combine domain-specific word embeddings with a label propagation framework to induce accurate domain-specific sentiment lexicons using small sets of seed words. We show that(More)