arXiv subjects

Robert West

Publications and source records attributed to Robert West.

At least 145 records · Page 8Linked to original sources

arXiv2021

Political Polarization in Online News Consumption

Political polarization appears to be on the rise, as measured by voting behavior, general affect towards opposing partisans and their parties, and contents posted and consumed online. Research over the years has focused on the role of the Web as a driver of polarization. In order to further our understanding of the factors behind online polarization, in the present work we collect and analyze Web browsing histories of tens of thousands of users alongside careful measurements of the time spent browsing various news sources. We show that online news consumption follows a polarized pattern, where users' visits to news sources aligned with their own political leaning are substantially longer than their visits to other news sources. Next, we show that such preferences hold at the individual as well as the population level, as evidenced by the emergence of clear partisan communities of news domains from aggregated browsing patterns. Finally, we tackle the important question of the role of user choices in polarization. Are users simply following the links proffered by their Web environment, or do they exacerbate partisan polarization by intentionally pursuing like-minded news sources? To answer this question, we compare browsing patterns with the underlying hyperlink structure spanned by the considered news domains, finding strong evidence of polarization in partisan browsing habits beyond that which can be explained by the hyperlink structure of the Web.

cs.SI↗

arXiv2021

YouNiverse: Large-Scale Channel and Video Metadata from English-Speaking YouTube

YouTube plays a key role in entertaining and informing people around the globe. However, studying the platform is difficult due to the lack of randomly sampled data and of systematic ways to query the platform's colossal catalog. In this paper, we present YouNiverse, a large collection of channel and video metadata from English-language YouTube. YouNiverse comprises metadata from over 136k channels and 72.9M videos published between May 2005 and October 2019, as well as channel-level time-series data with weekly subscriber and view counts. Leveraging channel ranks from socialblade.com, an online service that provides information about YouTube, we are able to assess and enhance the representativeness of the sample of channels. Additionally, the dataset also contains a table specifying which videos a set of 449M anonymous users commented on. YouNiverse, publicly available at https://doi.org/10.5281/zenodo.4650046, will empower the community to do research with and about YouTube.

cs.SI↗

arXiv2021

Evolution of Retweet Rates in Twitter User Careers: Analysis and Model

We study the evolution of the number of retweets received by Twitter users over the course of their "careers" on the platform. We find that on average the number of retweets received by users tends to increase over time. This is partly expected because users tend to gradually accumulate followers. Normalizing by the number of followers, however, reveals that the relative, per-follower retweet rate tends to be non-monotonic, maximized at a "peak age" after which it does not increase, or even decreases. We develop a simple mathematical model of the process behind this phenomenon, which assumes a constantly growing number of followers, each of whom loses interest over time. We show that this model is sufficient to explain the non-monotonic nature of per-follower retweet rates, without any assumptions about the quality of content posted at different times.

cs.SI↗

arXiv2021

Detection of an Atmosphere on a Rocky Exoplanet

We report the detection of an atmosphere on a rocky exoplanet, GJ 1132 b, which is similar to Earth in terms of size and density. The atmospheric transmission spectrum was detected using Hubble WFC3 measurements and shows spectral signatures of aerosol scattering, HCN, and CH$_{4}$ in a low mean molecular weight atmosphere. We model the atmospheric loss process and conclude that GJ 1132 b likely lost the original H/He envelope, suggesting that the atmosphere that we detect has been reestablished. We explore the possibility of H$_{2}$ mantle degassing, previously identified as a possibility for this planet by theoretical studies, and find that outgassing from ultrareduced magma could produce the observed atmosphere. In this way we use the observed exoplanet transmission spectrum to gain insights into magma composition for a terrestrial planet. The detection of an atmosphere on this rocky planet raises the possibility that the numerous powerfully irradiated Super-Earth planets, believed to be the evaporated cores of Sub-Neptunes, may, under favorable circumstances, host detectable atmospheres.

astro-ph.EP↗

arXiv2021

A pole to pole pressure temperature map of Saturn's thermosphere from Cassini Grand Finale data

Temperatures of the outer planet thermospheres exceed those predicted by solar heating alone by several hundred degrees. Enough energy is deposited at auroral regions to heat the entire thermosphere, but models predict that equatorward distribution is inhibited by strong Coriolis forces and ion drag. A better understanding of auroral energy deposition and circulation are critical to solving this so-called energy crisis. Stellar occultations observed by the Ultraviolet Imaging Spectrograph instrument during the Cassini Grand Finale were designed to map the thermosphere from pole to pole. We analyze these observations, together with earlier observations from 2016 and 2017, to create a two-dimensional map of densities and temperatures in Saturns thermosphere as a function of latitude and depth. The observed temperatures at auroral latitudes are cooler and peak at higher altitudes and lower latitudes than predicted by models, leading to a shallower meridional pressure gradient. Under modified geostrophy, we infer slower westward zonal winds that extend to lower latitudes than predicted, supporting equatorward flow from approximately 70 to 30 degrees latitude in both hemispheres. We also show evidence of atmospheric waves in the data that can contribute to equatorward redistribution of energy through zonal drag.

astro-ph.EP↗

arXiv2021

Formation of Social Ties Influences Food Choice: A Campus-Wide Longitudinal Study

Nutrition is a key determinant of long-term health, and social influence has long been theorized to be a key determinant of nutrition. It has been difficult to quantify the postulated role of social influence on nutrition using traditional methods such as surveys, due to the typically small scale and short duration of studies. To overcome these limitations, we leverage a novel source of data: logs of 38 million food purchases made over an 8-year period on the Ecole Polytechnique Federale de Lausanne (EPFL) university campus, linked to anonymized individuals via the smartcards used to make on-campus purchases. In a longitudinal observational study, we ask: How is a person's food choice affected by eating with someone else whose own food choice is healthy vs. unhealthy? To estimate causal effects from the passively observed log data, we control confounds in a matched quasi-experimental design: we identify focal users who at first do not have any regular eating partners but then start eating with a fixed partner regularly, and we match focal users into comparison pairs such that paired users are nearly identical with respect to covariates measured before acquiring the partner, where the two focal users' new eating partners diverge in the healthiness of their respective food choice. A difference-in-differences analysis of the paired data yields clear evidence of social influence: focal users acquiring a healthy-eating partner change their habits significantly more toward healthy foods than focal users acquiring an unhealthy-eating partner. We further identify foods whose purchase frequency is impacted significantly by the eating partner's healthiness of food choice. Beyond the main results, the work demonstrates the utility of passively sensed food purchase logs for deriving insights, with the potential of informing the design of public health interventions and food offerings.

cs.SI↗

arXiv2021

On the Value of Wikipedia as a Gateway to the Web

By linking to external websites, Wikipedia can act as a gateway to the Web. To date, however, little is known about the amount of traffic generated by Wikipedia's external links. We fill this gap in a detailed analysis of usage logs gathered from Wikipedia users' client devices. Our analysis proceeds in three steps: First, we quantify the level of engagement with external links, finding that, in one month, English Wikipedia generated 43M clicks to external websites, in roughly even parts via links in infoboxes, cited references, and article bodies. Official links listed in infoboxes have by far the highest click-through rate (CTR), 2.47% on average. In particular, official links associated with articles about businesses, educational institutions, and websites have the highest CTR, whereas official links associated with articles about geographical content, television, and music have the lowest CTR. Second, we investigate patterns of engagement with external links, finding that Wikipedia frequently serves as a stepping stone between search engines and third-party websites, effectively fulfilling information needs that search engines do not meet. Third, we quantify the hypothetical economic value of the clicks received by external websites from English Wikipedia, by estimating that the respective website owners would need to pay a total of $7--13 million per month to obtain the same volume of traffic via sponsored search. Overall, these findings shed light on Wikipedia's role not only as an important source of information, but also as a high-traffic gateway to the broader Web ecosystem.

cs.CY↗

arXiv2021

Crosslingual Topic Modeling with WikiPDA

We present Wikipedia-based Polyglot Dirichlet Allocation (WikiPDA), a crosslingual topic model that learns to represent Wikipedia articles written in any language as distributions over a common set of language-independent topics. It leverages the fact that Wikipedia articles link to each other and are mapped to concepts in the Wikidata knowledge base, such that, when represented as bags of links, articles are inherently language-independent. WikiPDA works in two steps, by first densifying bags of links using matrix completion and then training a standard monolingual topic model. A human evaluation shows that WikiPDA produces more coherent topics than monolingual text-based LDA, thus offering crosslinguality at no cost. We demonstrate WikiPDA's utility in two applications: a study of topical biases in 28 Wikipedia editions, and crosslingual supervised classification. Finally, we highlight WikiPDA's capacity for zero-shot language transfer, where a model is reused for new languages without any fine-tuning. Researchers can benefit from WikiPDA as a practical tool for studying Wikipedia's content across its 299 language editions in interpretable ways, via an easy-to-use library publicly available at https://github.com/epfl-dlab/WikiPDA.

cs.CL↗

arXiv2021

Calibration of Google Trends Time Series

Google Trends is a tool that allows researchers to analyze the popularity of Google search queries across time and space. In a single request, users can obtain time series for up to 5 queries on a common scale, normalized to the range from 0 to 100 and rounded to integer precision. Despite the overall value of Google Trends, rounding causes major problems, to the extent that entirely uninformative, all-zero time series may be returned for unpopular queries when requested together with more popular queries. We address this issue by proposing Google Trends Anchor Bank (G-TAB), an efficient solution for the calibration of Google Trends data. Our method expresses the popularity of an arbitrary number of queries on a common scale without being compromised by rounding errors. The method proceeds in two phases. In the offline preprocessing phase, an "anchor bank" is constructed, a set of queries spanning the full spectrum of popularity, all calibrated against a common reference query by carefully chaining together multiple Google Trends requests. In the online deployment phase, any given search query is calibrated by performing an efficient binary search in the anchor bank. Each search step requires one Google Trends request, but few steps suffice, as we demonstrate in an empirical evaluation. We make our code publicly available as an easy-to-use library at https://github.com/epfl-dlab/GoogleTrendsAnchorBank.

cs.SI↗

arXiv2020

Darks and Stripes: Effects of Clothing on Weight Perception

In many societies, appearing slim is considered attractive. The fashion industry has been attempting to cater to this trend by designing outfits that can enhance the appearance of slimness. Two anecdotal rules, widespread in the world of fashion, are (1) choose dark clothes and (2) avoid horizontal stripes, in order to appear slim. Thus far, empirical evidence has been unable to conclusively determine the validity of these rules, and there is consequently much controversy regarding the impact of both color and patterns on the visual perception of weight. In this paper, we aim to close this gap by presenting the results from a series of large-scale crowdsourcing studies that investigate the above two claims. We gathered a dataset of around 1,000 images of people from the Web together with their ground-truth weight and height, as well as clothing attributes about colors and patterns. To elicit the effects of colors and patterns, we asked crowd workers to estimate the weight in each image. For the analysis, we controlled potential confounds by matching images in pairs where the two images differ with respect to color or pattern, but are similar with respect to other relevant aspects. We created image pairs in two ways: first, observationally, i.e., from two real images; and second, experimentally, by manipulating the color or pattern of clothing in a real image via photo editing. Based on our analysis, we conclude that (1) dark clothes indeed decrease perceived weight slightly but statistically significantly, and (2) horizontal stripes have no discernible effect compared to solid light-colored clothes. These results contribute to advancing the debate around the effect of specific clothing colors and patterns and thus provide empirical grounds for everyday fashion decisions. Moreover, our work gives an outlook on the vast opportunities of using crowd sourcing in the modern fashion industry.

cs.CY↗

arXiv2020

Wikipedia Citations: A comprehensive dataset of citations with identifiers extracted from English Wikipedia

Wikipedia's contents are based on reliable and published sources. To this date, relatively little is known about what sources Wikipedia relies on, in part because extracting citations and identifying cited sources is challenging. To close this gap, we release Wikipedia Citations, a comprehensive dataset of citations extracted from Wikipedia. A total of 29.3M citations were extracted from 6.1M English Wikipedia articles as of May 2020, and classified as being to books, journal articles or Web contents. We were thus able to extract 4.0M citations to scholarly publications with known identifiers -- including DOI, PMC, PMID, and ISBN -- and further equip an extra 261K citations with DOIs from Crossref. As a result, we find that 6.7% of Wikipedia articles cite at least one journal article with an associated DOI, and that Wikipedia cites just 2% of all articles with a DOI currently indexed in the Web of Science. We release our code to allow the community to extend upon our work and update the dataset in the future.

cs.DL↗

arXiv2020

Node Attribute Completion in Knowledge Graphs with Multi-Relational Propagation

The existing literature on knowledge graph completion mostly focuses on the link prediction task. However, knowledge graphs have an additional incompleteness problem: their nodes possess numerical attributes, whose values are often missing. Our approach, denoted as MrAP, imputes the values of missing attributes by propagating information across the multi-relational structure of a knowledge graph. It employs regression functions for predicting one node attribute from another depending on the relationship between the nodes and the type of the attributes. The propagation mechanism operates iteratively in a message passing scheme that collects the predictions at every iteration and updates the value of the node attributes. Experiments over two benchmark datasets show the effectiveness of our approach.

cs.LG↗

arXiv2020

Human biases in body measurement estimation

Body measurements, including weight and height, are key indicators of health. Being able to visually assess body measurements reliably is a step towards increased awareness of overweight and obesity and is thus important for public health. Nevertheless it is currently not well understood how accurately humans can assess weight and height from images, and when and how they fail. To bridge this gap, we start from 1,682 images of persons collected from the Web, each annotated with the true weight and height, and ask crowd workers to estimate the weight and height for each image. We conduct a faceted analysis taking into account characteristics of the images as well as the crowd workers assessing the images, revealing several novel findings: (1) Even after aggregation, the crowd's accuracy is overall low. (2) We find strong evidence of contraction bias toward a reference value, such that the weight (height) of light (short) people is overestimated, whereas that of heavy (tall) people is underestimated. (3) We estimate workers' individual reference values using a Bayesian model, finding that reference values strongly correlate with workers' own height and weight, indicating that workers are better at estimating people similar to themselves. (4) The weight of tall people is underestimated more than that of short people; yet, knowing the height decreases the weight error only mildly. (5) Accuracy is higher on images of females than of males, but female and male workers are no different in terms of accuracy. (6) Crowd workers improve over time if given feedback on previous guesses. Finally, we explore various bias correction models for improving the crowd's accuracy, but find that this only leads to modest gains. Overall, this work provides important insights on biases in body measurement estimation as obesity related conditions are on the rise.

cs.SI↗

arXiv2020

KLearn: Background Knowledge Inference from Summarization Data

The goal of text summarization is to compress documents to the relevant information while excluding background information already known to the receiver. So far, summarization researchers have given considerably more attention to relevance than to background knowledge. In contrast, this work puts background knowledge in the foreground. Building on the realization that the choices made by human summarizers and annotators contain implicit information about their background knowledge, we develop and compare techniques for inferring background knowledge from summarization data. Based on this framework, we define summary scoring functions that explicitly model background knowledge, and show that these scoring functions fit human judgments significantly better than baselines. We illustrate some of the many potential applications of our framework. First, we provide insights into human information importance priors. Second, we demonstrate that averaging the background knowledge of multiple, potentially biased annotators or corpora greatly improves summary-scoring performance. Finally, we discuss potential applications of our framework beyond summarization.

cs.CL↗

arXiv2020

Adoption of Twitter's New Length Limit: Is 280 the New 140?

In November 2017, Twitter doubled the maximum allowed tweet length from 140 to 280 characters, a drastic switch on one of the world's most influential social media platforms. In the first long-term study of how the new length limit was adopted by Twitter users, we ask: Does the effect of the new length limit resemble that of the old one? Or did the doubling of the limit fundamentally change how Twitter is shaped by the limited length of posted content? By analyzing Twitter's publicly available 1% sample over a period of around 3 years, we find that, when the length limit was raised from 140 to 280 characters, the prevalence of tweets around 140 characters dropped immediately, while the prevalence of tweets around 280 characters rose steadily for about 6 months. Despite this rise, tweets approaching the length limit have been far less frequent after than before the switch. We find widely different adoption rates across languages and client-device types. The prevalence of tweets around 140 characters before the switch in a given language is strongly correlated with the prevalence of tweets around 280 characters after the switch in the same language, and very long tweets are vastly more popular on Web clients than on mobile clients. Moreover, tweets of around 280 characters after the switch are syntactically and semantically similar to tweets of around 140 characters before the switch, manifesting patterns of message squeezing in both cases. Taken together, these findings suggest that the new 280-character limit constitutes a new, less intrusive version of the old 140-character limit. The length limit remains an important factor that should be considered in all studies using Twitter data.

cs.SI↗

arXiv2020

Privacy-Preserving Classification with Secret Vector Machines

Today, large amounts of valuable data are distributed among millions of user-held devices, such as personal computers, phones, or Internet-of-things devices. Many companies collect such data with the goal of using it for training machine learning models allowing them to improve their services. User-held data is, however, often sensitive, and collecting it is problematic in terms of privacy. We address this issue by proposing a novel way of training a supervised classifier in a distributed setting akin to the recently proposed federated learning paradigm, but under the stricter privacy requirement that the server that trains the model is assumed to be untrusted and potentially malicious. We thus preserve user privacy by design, rather than by trust. In particular, our framework, called secret vector machine (SecVM), provides an algorithm for training linear support vector machines (SVM) in a setting in which data-holding clients communicate with an untrusted server by exchanging messages designed to not reveal any personally identifiable information. We evaluate our model in two ways. First, in an offline evaluation, we train SecVM to predict user gender from tweets, showing that we can preserve user privacy without sacrificing classification performance. Second, we implement SecVM's distributed framework for the Cliqz web browser and deploy it for predicting user gender in a large-scale online evaluation with thousands of clients, outperforming baselines by a large margin and thus showcasing that SecVM is suitable for production environments.

cs.LG↗

arXiv2020

Experts and authorities receive disproportionate attention on Twitter during the COVID-19 crisis

Timely access to accurate information is crucial during the COVID-19 pandemic. Prompted by key stakeholders' cautioning against an "infodemic", we study information sharing on Twitter from January through May 2020. We observe an overall surge in the volume of general as well as COVID-19-related tweets around peak lockdown in March/April 2020. With respect to engagement (retweets and likes), accounts related to healthcare, science, government and politics received by far the largest boosts, whereas accounts related to religion and sports saw a relative decrease in engagement. While the threat of an "infodemic" remains, our results show that social media also provide a platform for experts and public authorities to be widely heard during a global crisis.

cs.SI↗

arXiv2020

Population Dynamics in a Changing Environment: Random versus Periodic Switching

Environmental changes greatly influence the evolution of populations. Here, we study the dynamics of a population of two strains, one growing slightly faster than the other, competing for resources in a time-varying binary environment modeled by a carrying capacity switching either randomly or periodically between states of abundance and scarcity. The population dynamics is characterized by demographic noise (birth and death events) coupled to a varying environment. We elucidate the similarities and differences of the evolution subject to a stochastically- and periodically-varying environment. Importantly, the population size distribution is generally found to be broader under intermediate and fast random switching than under periodic variations, which results in markedly different asymptotic behaviors between the fixation probability of random and periodic switching. We also determine the detailed conditions under which the fixation probability of the slow strain is maximal.

q-bio.PE↗