<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:atom="http://www.w3.org/2005/Atom" version="2.0"><channel><title>Visualization Group Papers</title><link>http://vis.stanford.edu/papers/</link><description>Research papers published by the Stanford Visualization Group</description><atom:link href="http://vis.stanford.edu/feeds/papers/" rel="self"></atom:link><language>en-us</language><lastBuildDate>Wed, 01 Jan 2014 00:00:00 -0600</lastBuildDate><item><title>Visual Embedding: A Model for Visualization</title><link>http://vis.stanford.edu/papers/visual-embedding</link><description>We propose visual embedding as a model for automatically generating and evaluating visualizations. A visual embedding is a function from data points to a space of visual primitives that measurably preserves structures in the data (domain) within the mapped perceptual space (range). Visual embedding can serve as both a generative and an evaluative model. We demonstrate its use with three examples: coloring of neural tracts, scatter plots with icons, and evaluation of alternative diffusion tensor glyphs. We discuss several techniques for generating visual embedding functions, including probabilistic graphical models for embedding within discrete visual spaces. We also describe two complementary approaches - crowdsourcing and visual product spaces - for  building visual spaces with associated perceptual distance measures. Finally, we present future research directions for further developing the visual embedding model.</description><pubDate>Wed, 01 Jan 2014 00:00:00 -0600</pubDate><guid>http://vis.stanford.edu/papers/visual-embedding</guid></item><item><title>invis: Exploring High-Dimensional RNA Sequences from In Vitro Selection</title><link>http://vis.stanford.edu/papers/invis</link><description>In vitro selection and evolution is a powerful method for discovering RNA molecules based on their binding and catalysis properties. It has important applications to the study of genetic variation and molecular evolution. However, the resulting RNA sequences form a large, high-dimensional space and biologists lack adequate tools to explore and interpret these sequences. We present invis, the first visual analysis tool to facilitate exploration of in vitro selection sequence spaces. invis introduces a novel configuration of coordinated views that enables simultaneous inspection of global projections of sequence data alongside local regions of selected dimensions and sequence clusters. It allows scientists to isolate related sequences for further data analysis, compare sequence populations over varying conditions, filter sequences based on their similarities, and visualize likely pathways of genetic evolution. User feedback indicates that invis enables effective exploration of in vitro RNA selection sequences.</description><pubDate>Sat, 12 Oct 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/invis</guid></item><item><title>Differentiating Language Usage through Topic Models</title><link>http://vis.stanford.edu/papers/topic-models-language-usage</link><description>Sociologists wishing to employ topic models in their research need a helpful guide that describes the variety of topic modeling procedures, their issues, and various means of resolving them so as to convincingly answer sociological questions. We present this overview by recounting a series of our prior collaborative projects that have employed and developed various forms of topic models to understand language differentiation in academe. With each project, we encountered a variety of model-specific issues concerning the validity of topics and their suitability to our data and research questions. We developed a variety of novel visualization techniques to make sense of topic-solutions and used a variety of techniques to validate our results. In addition, we created a variety of new topic modeling techniques and procedures suitable to different kinds of data and research questions.</description><pubDate>Mon, 09 Sep 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/topic-models-language-usage</guid></item><item><title>Topic Model Diagnostics: Assessing Domain Relevance via Topical Alignment</title><link>http://vis.stanford.edu/papers/topic-model-diagnostics</link><description>The use of topic models to analyze domain-specific texts often requires manual validation of the latent topics to ensure that they are meaningful. We introduce a framework to support such a large-scale assessment of topical relevance. We measure the correspondence between a set of latent topics and a set of reference concepts to quantify four types of topical misalignment: junk, fused, missing, and repeated topics. Our analysis compares 10,000 topic model variants to 200 expert-provided domain concepts, and demonstrates how our framework can inform choices of model parameters, inference algorithms, and intrinsic measures of topical quality.</description><pubDate>Sun, 14 Jul 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/topic-model-diagnostics</guid></item><item><title>imMens: Real-time Visual Querying of Big Data</title><link>http://vis.stanford.edu/papers/immens</link><description>Data analysts must make sense of increasingly large data sets, sometimes with billions or more records. We present methods for interactive visualization of big data, following the principle that perceptual and interactive scalability should be limited by the chosen resolution of the visualized data, not the number of records. We first describe a design space of scalable visual summaries that use data reduction methods (such as binned aggregation or sampling) to visualize a variety of data types. We then contribute methods for interactive querying (e.g., brushing &amp;amp; linking) among binned plots through a combination of multivariate data tiles and parallel query processing. We implement our techniques in imMens, a browser-based visual analysis system that uses WebGL for data processing and rendering on the GPU. In benchmarks imMens sustains 50 frames-per-second brushing &amp;amp; linking among dozens of visualizations, with invariant performance on data sizes ranging from thousands to billions of records.</description><pubDate>Sun, 02 Jun 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/immens</guid></item><item><title>Selecting Semantically-Resonant Colors for Data Visualization</title><link>http://vis.stanford.edu/papers/semantically-resonant-colors</link><description>We introduce an algorithm for automatic selection of semantically-resonant colors to represent data (e.g., using blue for data about &amp;quot;oceans&amp;quot;, or pink for &amp;quot;love&amp;quot;). Given a set of categorical values and a target color palette, our algorithm matches each data value with a unique color. Values are mapped to colors by collecting representative images, analyzing image color distributions to determine value-color affinity scores, and choosing an optimal assignment. Our affinity score balances the probability of a color with how well it discriminates among data values. A controlled study shows that expert-chosen semantically-resonant colors improve speed on chart reading tasks compared to a standard palette, and that our algorithm selects colors that lead to similar gains. A second study verifies that our algorithm effectively selects colors across a variety of data categories.</description><pubDate>Sat, 01 Jun 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/semantically-resonant-colors</guid></item><item><title>Identifying Medical Terms in Patient-Authored Text: A Crowdsourcing-based Approach</title><link>http://vis.stanford.edu/papers/adept</link><description>Background and objective: As people increasingly engage in online health-seeking behavior and contribute to health-oriented websites, the volume of medical text authored by patients and other medical novices grows rapidly. However, we lack an effective method for automatically identifying medical terms in patient-authored text (PAT). We demonstrate that crowdsourcing PAT medical term identification tasks to non-experts is a viable method for creating large, accurately-labeled PAT datasets; moreover, such datasets can be used to train classifiers that outperform existing medical term identification tools.

Materials and methods: To evaluate the viability of using non-expert crowds to label PAT, we compare expert (registered nurses) and non-expert (Amazon Mechanical Turk workers; Turkers) responses to a PAT medical term identification task. Next, we build a crowd-labeled dataset comprising 10 000 sentences from MedHelp. We train two models on this dataset and evaluate their performance, as well as that of MetaMap, Open Biomedical Annotator (OBA), and NaCTeM&amp;#39;s TerMINE, against two gold standard datasets: one from MedHelp and the other from CureTogether.

Results: When aggregated according to a corroborative voting policy, Turker responses predict expert responses with an F1 score of 84%. A conditional random field (CRF) trained on 10 000 crowd-labeled MedHelp sentences achieves an F1 score of 78% against the CureTogether gold standard, widely outperforming OBA (47%), TerMINE (43%), and MetaMap (39%). A failure analysis of the CRF suggests that misclassified terms are likely to be either generic or rare.

Conclusions: Our results show that combining statistical models sensitive to sentence-level context with crowd-labeled data is a scalable and effective technique for automatically identifying medical terms in PAT.</description><pubDate>Mon, 06 May 2013 00:00:00 -0500</pubDate><guid>http://vis.stanford.edu/papers/adept</guid></item><item><title>Colony Life History and Lifetime Reproductive Success of Red Harvester Ant Colonies</title><link>http://vis.stanford.edu/papers/colony-life-history</link><description>We estimate colony reproductive success, in numbers of offspring colonies arising from a colony&amp;#39;s daughter queens, of colonies of the red harvester ant, Pogonomyrmex barbatus. A measure of lifetime reproductive success is essential to understand the relation of ecological factors, phenotype and fitness in a natural population. This was possible for the first time in a natural population of ant colonies using data from long-term study of a population of colonies in south-eastern Arizona, for which ages of all colonies are known from census data collected since 1985. Parentage analyses of microsatellite data from 5 highly polymorphic loci were used to assign offspring colonies to maternal parent colonies in a population of about 265 colonies, ages 1–28 years, sampled in 2010. The estimated population growth rate Ro was 1-69 and generation time was 7-8 years. There was considerable variation among colonies in reproductive success: of 199 possible parent colonies, only 49 (~25%) had offspring colonies on the site. The mean number of offspring colonies per maternal parent colony was 2·94 and ranged from 1 to 8. A parent was identified for the queen of 146 of 247 offspring colonies. There was no evidence for reproductive senescence; fecundity was about the same throughout the 25–30 year lifespan of a colony. There were no trends in the distance or direction of the dispersal of an offspring relative to its maternal parent colony. There was no relationship between the number of gynes produced by a colony in 1 year and the number of offspring colonies subsequently founded by its daughter reproductive females. The results provide the first estimate of a life table for a population of ant colonies and the first estimate of the female component of colony lifetime reproductive success. The results suggest that commonly used measures of reproductive output may not be correlated with realized reproductive success. This is the starting point for future investigation asking whether variation in reproductive success is related to phenotypic variation among colonies in behavioural and ecological traits.</description><pubDate>Fri, 01 Feb 2013 00:00:00 -0600</pubDate><guid>http://vis.stanford.edu/papers/colony-life-history</guid></item><item><title>The Efﬁcacy of Human Post-Editing for Language Translation</title><link>http://vis.stanford.edu/papers/post-editing</link><description>Language translation is slow and expensive, so various forms of machine assistance have been devised. Automatic machine translation systems process text quickly and cheaply, but with quality far below that of skilled human translators. To bridge this quality gap, the translation industry has investigated post-editing, or the manual correction of machine output. We present the first rigorous, controlled analysis of post-editing and find that post-editing leads to reduced time and, surprisingly, improved quality for three diverse language pairs (English to Arabic, French, and German). Our statistical models and visualizations of experimental data indicate that some simple predictors (like source text part of speech counts) predict translation time, and that post-editing results in very different interaction patterns. From these results we distill implications for the design of new language translation interfaces.</description><pubDate>Mon, 21 Jan 2013 00:00:00 -0600</pubDate><guid>http://vis.stanford.edu/papers/post-editing</guid></item><item><title>Webzeitgeist: Design Mining the Web</title><link>http://vis.stanford.edu/papers/webzeitgeist</link><description>Advances in data mining and knowledge discovery have transformed the way Web sites are designed. However, while visual presentation is an intrinsic part of the Web, traditional data mining techniques ignore render-time page structures and their attributes. This paper introduces design mining for the Web: using knowledge discovery techniques to understand design demographics, automate design curation, and support data-driven design tools. This idea is manifest in Webzeitgeist, a platform for large-scale design mining comprising a repository of over 100,000 Web pages and 100 million design elements. This paper describes the principles driving design mining, the implementation of the Webzeitgeist architecture, and the new class of data-driven design applications it enables.</description><pubDate>Sun, 20 Jan 2013 00:00:00 -0600</pubDate><guid>http://vis.stanford.edu/papers/webzeitgeist</guid></item></channel></rss>