AttentionViz: A Global View of Transformer Attention.

Transformer models are revolutionizing machine learning, but their inner workings remain mysterious. In this work, we present a new visualization technique designed to help researchers understand the self-attention mechanism in transformers that allows these models to learn rich, contextual relationships between elements of a sequence. The main idea behind our method is to visualize a joint embedding of the query and key vectors used by transformer models to compute attention. Unlike previous attention visualization techniques, our approach enables the analysis of global patterns across multiple input sequences. We create an interactive visualization tool, AttentionViz (demo: http://attentionviz.com), based on these joint query-key embeddings, and use it to study attention mechanisms in both language and vision transformers. We demonstrate the utility of our approach in improving model understanding and offering new insights about query-key interactions through several application scenarios and expert feedback.

Grand Challenges in Visual Analytics Applications.

In the past two decades, research in visual analytics (VA) applications has made tremendous progress, not just in terms of scientific contributions, but also in real-world impact across wide-ranging domains including bioinformatics, urban analytics, and explainable AI. Despite these success stories, questions on the rigor and value of VA application research have emerged as a grand challenge. This article outlines a research and development agenda for making VA application research more rigorous and impactful. We first analyze the characteristics of VA application research and explain how they cause the rigor and value problem. Next, we propose a research ecosystem for improving scientific value, and rigor and outline an agenda with 12 open challenges spanning four areas, including foundation, methodology, application, and community. We encourage discussions, debates, and innovative efforts toward more rigorous and impactful VA research.

The sleep as a predictor of musculoskeletal injuries in adolescent athletes.

Objectives: Sleep is essential for musculoskeletal and cognitive recovery. Adolescent athletes tend to sleep poorly compared to adults and it may predispose them to sports injuries. Our aims are to estimate whether the quantity/quality of sleep are associated with sports injuries in adolescent athletes and to compare the quantity/quality of sleep between the training and competition seasons, and the school vacation period. Material and Methods: It was a cohort study with 19 track and field athletes of both sexes, aged between 12 and 21 years. We evaluated their sleep-wake habit through actigraphy during three phases: 1 - mid-season, 2 - competition, and 3 - school vacation. The previous six months injury history and the occurrence of injuries in a six-month follow-up were recorded. Logistic regression and variance analysis were performed. The significance level used was 0.05. Results: Wake after sleep onset (WASO) predicted previous injuries (OR=1.144) and time awake (TA) predicted injury occurrence (OR=0.974). TA decreased from phase 2 to phase 3 (p=0.004), total sleep time (TST) increased from phase 2 to phase 3 (p=0.012), and WASO decreased between phases 1 and 2 (p=0.001) and between phases 1 and 3 (p=0.025). Conclusion: Our study demonstrated that the quantity and quality of sleep were associated with musculoskeletal injuries in adolescent track and field athletes. Previous injuries were predicted by WASO and the occurrence of injuries was predicted by TA. Furthermore, during the vacation period they had lower TA and WASO, and higher TST than on school days.

Concurrent Validation and Reference Values of Gluteus Medius Clinical Test.

CONTEXT: The hip abductor muscles, mainly the gluteus medius, are responsible for controlling hip adduction in a closed kinetic chain. Frontal plane knee alignment, assessed during functional activities such squatting, jumping and running, may overload joint structures, like the anterior cruciate ligament and patellofemoral joint. The hand-held dynamometer is reliable and effective for testing the muscular strength of the hip abductors. OBJECTIVES: 1. To assess the concurrent validity between the gluteus medius clinical test and a maximum isometric force test of the hip abductors using the hand-held dynamometer; (2) to determine the intra and inter-examiner reliability for the application of the gluteus medius clinical test; and (3) to describe reference values of gluteus medius clinical test on a population of youth athletes. DESIGN: Cross-sectional. METHODS: Thirty healthy individuals were recruited for validity and reliability testing. On the first day, participants performed the maximal isometric test of the hip abductors, measured via hand-held dynamometry. On the following week, the gluteus medius clinical test was performed. Intraclass correlation coefficients (ICC2,2) were computed for the reliability analysis, with a 95% confidence interval. To generate reference values, the gluteus medius clinical test was performed on 273 athletes. RESULTS: The results of this study indicated a weak positive correlation (r = 0.436, p = 0.001) between tests, which indicates that they examine different domains of gluteus medius muscle function, likely endurance and muscle strength. The magnitude of computed ICCs (>0.95) indicates excellent intra- and inter-examiner reliability. CONCLUSION: The findings of the current study indicate that the gluteus medius clinical test is reliable and examines a domain of muscular function not fully captured by HHD. The clinical test developed in this study is low-cost and can be included for gluteus medius assessment. LEVEL OF EVIDENCE: Level 3.

The What-If Tool: Interactive Probing of Machine Learning Models.

A key challenge in developing and deploying Machine Learning (ML) systems is understanding their performance across a wide range of inputs. To address this challenge, we created the What-If Tool, an open-source application that allows practitioners to probe, visualize, and analyze ML systems, with minimal coding. The What-If Tool lets practitioners test performance in hypothetical situations, analyze the importance of different data features, and visualize model behavior across multiple models and subsets of input data. It also lets practitioners measure systems according to multiple ML fairness metrics. We describe the design of the tool, and report on real-life usage at different organizations.

Poor Sleep Quality's Association With Soccer Injuries: Preliminary Data.

PURPOSE: To investigate the relationship between sleep quality and quantity and injuries in elite soccer players and to compare sleep-wake variables and injury characteristics. METHODS: The current investigation was a prospective cohort study of 23 elite male soccer players competing for 2 teams over 6 mo in the highest-level Brazilian competition. The players' sleep behavior was monitored for 10 d in the preseason using self-reporting sleep diaries and wrist activity monitors to determine sleep duration and quality. Furthermore, injuries were recorded by the respective club's medical teams into a specific database. Details of injuries recorded included the type, location, and severity of each injury. The results were expressed as descriptive statistics, and the significance level was set at 5%. The Mann-Whitney U test was performed to compare the sleep variables between groups. Spearman correlation coefficient and linear-regression analysis were used. RESULTS: The results indicated a moderate negative correlation between sleep efficiency and particular injury characteristics, including absence time, injury severity, and amount of injuries. The linear-regression analysis indicated that 44% of the total variance in the number of injuries can be explained by sleep efficiency, 24% of the total variance in the absence time after injury (days) can be explained by sleep efficiency, and 47% of the total variance in the injury severity can be explained by sleep efficiency. CONCLUSIONS: Soccer players who exhibit lower sleep quality or nonrestorative sleep show associations with increased number and severity of musculoskeletal injuries.

GAN Lab: Understanding Complex Deep Generative Models using Interactive Visual Experimentation.

Recent success in deep learning has generated immense interest among practitioners and students, inspiring many to learn about this new technology. While visual and interactive approaches have been successfully developed to help people more easily learn deep learning, most existing tools focus on simpler models. In this work, we present GAN Lab, the first interactive visualization tool designed for non-experts to learn and experiment with Generative Adversarial Networks (GANs), a popular class of complex deep learning models. With GAN Lab, users can interactively train generative models and visualize the dynamic training process's intermediate results. GAN Lab tightly integrates an model overview graph that summarizes GAN's structure, and a layered distributions view that helps users interpret the interplay between submodels. GAN Lab introduces new interactive experimentation features for learning complex deep learning models, such as step-by-step training at multiple levels of abstraction for understanding intricate training dynamics. Implemented using TensorFlow.js, GAN Lab is accessible to anyone via modern web browsers, without the need for installation or specialized hardware, overcoming a major practical challenge in deploying interactive tools for deep learning.

Deep learning of aftershock patterns following large earthquakes.

Aftershocks are a response to changes in stress generated by large earthquakes and represent the most common observations of the triggering of earthquakes. The maximum magnitude of aftershocks and their temporal decay are well described by empirical laws (such as Bath's law1 and Omori's law2), but explaining and forecasting the spatial distribution of aftershocks is more difficult. Coulomb failure stress change3 is perhaps the most widely used criterion to explain the spatial distributions of aftershocks4-8, but its applicability has been disputed9-11. Here we use a deep-learning approach to identify a static-stress-based criterion that forecasts aftershock locations without prior assumptions about fault orientation. We show that a neural network trained on more than 131,000 mainshock-aftershock pairs can predict the locations of aftershocks in an independent test dataset of more than 30,000 mainshock-aftershock pairs more accurately (area under curve of 0.849) than can classic Coulomb failure stress change (area under curve of 0.583). We find that the learned aftershock pattern is physically interpretable: the maximum change in shear stress, the von Mises yield criterion (a scaled version of the second invariant of the deviatoric stress-change tensor) and the sum of the absolute values of the independent components of the stress-change tensor each explain more than 98 per cent of the variance in the neural-network prediction. This machine-learning-driven insight provides improved forecasts of aftershock locations and identifies physical quantities that may control earthquake triggering during the most active part of the seismic cycle.

Visualizing Dataflow Graphs of Deep Learning Models in TensorFlow.

We present a design study of the TensorFlow Graph Visualizer, part of the TensorFlow machine intelligence platform. This tool helps users understand complex machine learning architectures by visualizing their underlying dataflow graphs. The tool works by applying a series of graph transformations that enable standard layout techniques to produce a legible interactive diagram. To declutter the graph, we decouple non-critical nodes from the layout. To provide an overview, we build a clustered graph using the hierarchical structure annotated in the source code. To support exploration of nested structure on demand, we perform edge bundling to enable stable and responsive cluster expansion. Finally, we detect and highlight repeated structures to emphasize a model's modular composition. To demonstrate the utility of the visualizer, we describe example usage scenarios and report user feedback. Overall, users find the visualizer useful for understanding, debugging, and sharing the structures of their models.

Participatory visualization with Wordle.

We discuss the design and usage of "Wordle," a web-based tool for visualizing text. Wordle creates tag-cloud-like displays that give careful attention to typography, color, and composition. We describe the algorithms used to balance various aesthetic criteria and create the distinctive Wordle layouts. We then present the results of a study of Wordle usage, based both on spontaneous behaviour observed in the wild, and on a large-scale survey of Wordle users. The results suggest that Wordles have become a kind of medium of expression, and that a "participatory culture" has arisen around them.

Mapping text with phrase nets.

We present a new technique, the phrase net, for generating visual overviews of unstructured text. A phrase net displays a graph whose nodes are words and whose edges indicate that two words are linked by a user-specified relation. These relations may be defined either at the syntactic or lexical level; different relations often produce very different perspectives on the same text. Taken together, these perspectives often provide an illuminating visual overview of the key concepts and relations in a document or set of documents.

The Word Tree, an interactive visual concordance.

We introduce the Word Tree, a new visualization and information-retrieval technique aimed at text documents. A word tree is a graphical version of the traditional "keyword-in-context" method, and enables rapid querying and exploration of bodies of text. In this paper we describe the design of the technique, along with some of the technical issues that arise in its implementation. In addition, we discuss the results of several months of public deployment of word trees on Many Eyes, which provides a window onto the ways in which users obtain value from the visualization.

ManyEyes: a site for visualization at internet scale.

We describe the design and deployment of Many Eyes, a public web site where users may upload data, create interactive visualizations, and carry on discussions. The goal of the site is to support collaboration around visualizations at a large scale by fostering a social style of data analysis in which visualizations not only serve as a discovery tool for individuals but also as a medium to spur discussion among users. To support this goal, the site includes novel mechanisms for end-user creation of visualizations and asynchronous collaboration around those visualizations. In addition to describing these technologies, we provide a preliminary report on the activity of our users.