arXiv subjects

James McInerney

Publications and source records attributed to James McInerney.

25 records · Page 2Linked to original sources

arXiv2016

Modeling User Exposure in Recommendation

Collaborative filtering analyzes user preferences for items (e.g., books, movies, restaurants, academic papers) by exploiting the similarity patterns across users. In implicit feedback settings, all the items, including the ones that a user did not consume, are taken into consideration. But this assumption does not accord with the common sense understanding that users have a limited scope and awareness of items. For example, a user might not have heard of a certain paper, or might live too far away from a restaurant to experience it. In the language of causal analysis, the assignment mechanism (i.e., the items that a user is exposed to) is a latent variable that may change for various user/item combinations. In this paper, we propose a new probabilistic approach that directly incorporates user exposure to items into collaborative filtering. The exposure is modeled as a latent variable and the model infers its value from data. In doing so, we recover one of the most successful state-of-the-art approaches as a special case of our model, and provide a plug-in method for conditioning exposure on various forms of exposure covariates (e.g., topics in text, venue locations). We show that our scalable inference algorithm outperforms existing benchmarks in four different domains both with and without exposure covariates.

stat.ML↗

arXiv2015

Cosmography of KNdS Black Holes and Isentropic Phase Transitions

We present a new analysis of Kerr-Newman-deSitter black holes in terms of thermodynamic quantities that are defined in the observable portion of the universe; between the black hole and cosmological horizons. In particular, we replace the mass $m$ with a new 'area product' parameter $X$. The physical region of parameter space is found analytically and thermodynamic quantities are given by simple algebraic functions of these parameters. We find that different geometrical properties of the black holes are usefully distinguished by the sum of the black hole and cosmological entropies. The physical parameter space breaks into a region in which the total entropy, together with $Λ$, $a$ and $q$ uniquely specifies the black hole, and a region in which there is a two-fold degeneracy. In this latter region, there are isentropic pairs of black holes, having the same $Λ$, $a$, and $q$, but different $X$. The thermodynamic volumes and masses differ in such that there are high and low density branches. The partner spacetimes are related by a simple inversion of $X$, which has a fixed point at the state of maximal total entropy. We compute the compressibility at fixed total entropy and find that it diverges at the maximal entropy point. Hence a picture emerges of high and low density phases merging at this critical point.

hep-th↗

arXiv2015

Dynamic Poisson Factorization

Models for recommender systems use latent factors to explain the preferences and behaviors of users with respect to a set of items (e.g., movies, books, academic papers). Typically, the latent factors are assumed to be static and, given these factors, the observed preferences and behaviors of users are assumed to be generated without order. These assumptions limit the explorative and predictive capabilities of such models, since users' interests and item popularity may evolve over time. To address this, we propose dPF, a dynamic matrix factorization model based on the recent Poisson factorization model for recommendations. dPF models the time evolving latent factors with a Kalman filter and the actions with Poisson distributions. We derive a scalable variational inference algorithm to infer the latent factors. Finally, we demonstrate dPF on 10 years of user click data from arXiv.org, one of the largest repository of scientific papers and a formidable source of information about the behavior of scientists. Empirically we show performance improvement over both static and, more recently proposed, dynamic recommendation models. We also provide a thorough exploration of the inferred posteriors over the latent variables.

cs.LG↗

arXiv2015

The Population Posterior and Bayesian Inference on Streams

Many modern data analysis problems involve inferences from streaming data. However, streaming data is not easily amenable to the standard probabilistic modeling approaches, which assume that we condition on finite data. We develop population variational Bayes, a new approach for using Bayesian modeling to analyze streams of data. It approximates a new type of distribution, the population posterior, which combines the notion of a population distribution of the data with Bayesian inference in a probabilistic model. We study our method with latent Dirichlet allocation and Dirichlet process mixtures on several large-scale data sets.

stat.ML↗

arXiv2014

Targeted Social Mobilisation in a Global Manhunt

Social mobilization, the ability to mobilize large numbers of people via social networks to achieve highly distributed tasks, has received significant attention in recent times. This growing capability, facilitated by modern communication technology, is highly relevant to endeavors which require the search for individuals that posses rare information or skill, such as finding medical doctors during disasters, or searching for missing people. An open question remains, as to whether in time-critical situations, people are able to recruit in a targeted manner, or whether they resort to so-called blind search, recruiting as many acquaintances as possible via broadcast communication. To explore this question, we examine data from our recent success in the U.S. State Department's Tag Challenge, which required locating and photographing 5 target persons in 5 different cities in the United States and Europe in less than 12 hours, based only on a single mug-shot. We find that people are able to consistently route information in a targeted fashion even under increasing time pressure. We derive an analytical model for global mobilization and use it to quantify the extent to which people were targeting others during recruitment. Our model estimates that approximately 1 in 3 messages were of targeted fashion during the most time-sensitive period of the challenge.This is a novel observation at such short temporal scales, and calls for opportunities for devising viral incentive schemes that provide distance- or time-sensitive rewards to approach the target geography more rapidly, with applications in multiple areas from emergency preparedness, to political mobilization.

physics.soc-ph↗

arXiv2013

Learning Periodic Human Behaviour Models from Sparse Data for Crowdsourcing Aid Delivery in Developing Countries

In many developing countries, half the population lives in rural locations, where access to essentials such as school materials, mosquito nets, and medical supplies is restricted. We propose an alternative method of distribution (to standard road delivery) in which the existing mobility habits of a local population are leveraged to deliver aid, which raises two technical challenges in the areas optimisation and learning. For optimisation, a standard Markov decision process applied to this problem is intractable, so we provide an exact formulation that takes advantage of the periodicities in human location behaviour. To learn such behaviour models from sparse data (i.e., cell tower observations), we develop a Bayesian model of human mobility. Using real cell tower data of the mobility behaviour of 50,000 individuals in Ivory Coast, we find that our model outperforms the state of the art approaches in mobility prediction by at least 25% (in held-out data likelihood). Furthermore, when incorporating mobility prediction with our MDP approach, we find a 81.3% reduction in total delivery time versus routine planning that minimises just the number of participants in the solution path.

cs.AI↗

arXiv2012

Horizontal gene transfer drives extreme physiological change in Haloarchaea

The haloarchaea are aerobic, heterotrophic, photophosphorylating prokaryotes, whose supposed closest relatives and ancestors, the methanogens, are CO2-reducing, anaerobic chemolithotrophs. Using two available haloarchaeal genomes we firstly confirmed the methanogenic ancestry of the group and then investigated those individual genes in the haloarchaea that differ in their phylogenetic signal to this relationship. We found that almost half the genes, about which we can make strong statements, have bacterial ancestry and are likely a result of multiple horizontal transfer events. Futhermore their functions specifically relate to the phenotypic changes required for a chemolithotroph to become a heterotroph. If this phylogenetic relationship is correct, it implies the development of the haloarchaeal phenotype was among the most extreme changes in cellular physiology fuelled by horizontal gene transfer.

q-bio.GN↗