A5039276358- Electoral Systems and Political Participation
- Data Visualization and Analytics
- Topic Modeling
- Social Media and Politics
- Multimedia Communication and Technology
- Natural Language Processing Techniques
- Explainable Artificial Intelligence (XAI)
- Video Analysis and Summarization
- Political and Economic history of UK and US
- Populism, Right-Wing Movements
- Social Policy and Reform Studies
- Youth Education and Societal Dynamics
- Political Influence and Corporate Strategies
- Data Management and Algorithms
- American Constitutional Law and Politics
- Ethics and Social Impacts of AI
- Silicon Carbide Semiconductor Technologies
- Advanced DC-DC Converters
- Advanced Text Analysis Techniques
- Aesthetic Perception and Analysis
- Neural Networks and Applications
- Software Engineering Research
- Wikis in Education and Collaboration
- Anomaly Detection Techniques and Applications
- Computer Graphics and Visualization Techniques
Google (United States)
2014-2024
Harvard University
2021-2024
Infineon Technologies (Austria)
2023-2024
University of Michigan
1981-2023
Harvard University Press
1993-2023
Reutlingen University
2016-2019
NTT (Japan)
2019
Robert Bosch (Germany)
2017
University of California, Irvine
1996-2016
University of Houston
2013
TensorFlow is an interface for expressing machine learning algorithms, and implementation executing such algorithms. A computation expressed using can be executed with little or no change on a wide variety of heterogeneous systems, ranging from mobile devices as phones tablets up to large-scale distributed systems hundreds machines thousands computational GPU cards. The system flexible used express including training inference algorithms deep neural network models, it has been conducting...
INTRODUCTION CHAPTER 1 UNTHINKABLE DEMOCRACY: POLITICAL CHANGE IN ADVANCED INDUSTRIAL DEMOCRACIES PART I PARTIES THE ELECTORATE 2 DECLINE OF PARTY IDENTIFICATIONS 3 CONSEQUENCES PARTISAN DEALIGNMENT 4 MOBILIZATION II AS ORGANIZATIONS 5 WITHOUT MEMBERS? ORGANIZATION A CHANGING ELECTORAL ENVIRONMENT 6 CAMPAIGN 7 FROM SOCIAL INTEGRATION TO CONTESTATION: DISTRIBUTION POWER WITHIN III GOVERNMENT 8 LEGISLATURE: TWO COMPETING EXPLANATIONS 9 AT CORE 10 PLATFORM DECLARATIONS POLICY OUTCOMES: PROFILES...
We propose a simple solution to use single Neural Machine Translation (NMT) model translate between multiple languages. Our requires no changes the architecture from standard NMT system but instead introduces an artificial token at beginning of input sentence specify required target language. Using shared wordpiece vocabulary, our approach enables Multilingual systems using model. On WMT’14 benchmarks, multilingual achieves comparable performance for English→French and surpasses...
We combine well-known techniques from the areas of error-correcting codes and cryptography to achieve a new type cryptographic primitive that we refer as fuzzy commitment scheme. Like conventional scheme, our scheme is both concealing binding: it infeasible for an attacker learn committed value, also committer decommit value in more than one way. In must be opened using unique witness, which acts, essentially, decryption key. By contrast, sense accepts witness close original encrypting...
Explaining the output of a deep network remains challenge. In case an image classifier, one type explanation is to identify pixels that strongly influence final decision. A starting point for this strategy gradient class score function with respect input image. This can be interpreted as sensitivity map, and there are several techniques elaborate on basic idea. paper makes two contributions: it introduces SmoothGrad, simple method help visually sharpen gradient-based maps, discusses lessons...
Predicting ad click-through rates (CTR) is a massive-scale learning problem that central to the multi-billion dollar online advertising industry. We present selection of case studies and topics drawn from recent experiments in setting deployed CTR prediction system. These include improvements context traditional supervised based on an FTRL-Proximal algorithm (which has excellent sparsity convergence properties) use per-coordinate rates.
The Internet has fostered an unconventional and powerful style of collaboration: "wiki" web sites, where every visitor the power to become editor. In this paper we investigate dynamics Wikipedia, a prominent, thriving wiki. We make three contributions. First, introduce new exploratory data analysis tool, history flow visualization, which is effective in revealing patterns within wiki context believe will be useful other collaborative situations as well. Second, discuss several collaboration...
The interpretation of deep learning models is a challenge due to their size, complexity, and often opaque internal state. In addition, many systems, such as image classifiers, operate on low-level features rather than high-level concepts. To address these challenges, we introduce Concept Activation Vectors (CAVs), which provide an neural net's state in terms human-friendly key idea view the high-dimensional net aid, not obstacle. We show how use CAVs part technique, Testing with (TCAV), that...
The purpose of this thesis is to evaluate unsupervised clustering strategies for data associated with bulk carrier trucks, in the detection failure patterns thermal systems.The study these techniques important field data-based predictive maintenance implementation machine learning algorithms that allow proper planning schedules freight transport companies.For development thesis, telemetry devices tractor trucks a Colombian cargo company are used as source information, which report real time...
We describe the design and deployment of Many Eyes, a public Web site where users may upload data, create interactive visualizations, carry on discussions. The goal is to support collaboration around visualizations at large scale by fostering social style data analysis in which not only serve as discovery tool for individuals but also medium spur discussion among users. To this goal, includes novel mechanisms end-user creation asynchronous those visualizations. In addition describing these...
This article applies theories of social cognition in an investigation the dimensions assessments candidates employed by voters United States. An empirical description public's cognitive representations presidential candidates, derived from responses to open-ended questions American National Election Studies 1952 1984, reveals that perceptions are generally focused on “personality” characteristics rather than issue concerns or partisan group connections. Contrary implications past research,...
Treemaps, a space-filling method for visualizing large hierarchical data sets, are receiving increasing attention. Several algorithms have been previously proposed to create more useful displays by controlling the aspect ratios of rectangles that make up treemap. While these do improve visibility small items in single layout, they introduce instability over time display dynamically changing data, fail preserve order underlying and layouts difficult visually search. In addition, continuous...
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, analyze ML systems, with minimal coding. The Tool lets test hypothetical situations, importance different data features, visualize model behavior multiple models subsets input data. It also measure according fairness metrics....
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, composition. describe algorithms used balance various aesthetic criteria create distinctive layouts. then present results study usage, based both on spontaneous behaviour observed in wild, large-scale survey users. The suggest Wordles have become kind medium expression, "participatory culture" has arisen around them.
Wikipedia, the online encyclopedia, has attracted attention both because of its popularity and unconventional policy letting anyone on Internet edit articles. This paper describes results an empirical analysis Wikipedia discusses ways in which community evolved as it grown. We contrast our findings with earlier study present three main results. First, maintains a strong resilience to malicious editing, despite tremendous growth high traffic. Second, fastest growing areas are devoted...
In February 2008, the New York Times published an unusual chart of box office revenues for 7500 movies over 21 years. The was based on a similar visualization, developed by first author, that displayed trends in music listening. This paper describes design decisions and algorithms behind these graphics, discusses reaction Web. We suggest this type complex layered graph is effective displaying large data sets to mass audience. provide mathematical analysis how relates traditional stacked...
Machine learning (ML) is increasingly being used in image retrieval systems for medical decision making. One application of ML to retrieve visually similar images from past patients (e.g. tissue biopsies) reference when making a with new patient. However, no algorithm can perfectly capture an expert's ideal notion similarity every case: that algorithmically determined be may not medically relevant doctor's specific diagnostic needs. In this paper, we identified the needs pathologists...
We present a design study of the TensorFlow Graph Visualizer, part machine intelligence platform. This tool helps users understand complex learning architectures by visualizing their underlying dataflow graphs. The works applying series graph transformations that enable standard layout techniques to produce legible interactive diagram. To declutter graph, we decouple non-critical nodes from layout. provide an overview, build clustered using hierarchical structure annotated in source code....
We introduce the Word Tree, a new visualization and information-retrieval technique aimed at text documents. A word tree is graphical version of traditional "keyword-in-context" method, enables rapid querying exploration bodies text. In this paper we describe design technique, along with some technical issues that arise in its implementation. addition, discuss results several months public deployment trees on Many Eyes, which provides window onto ways users obtain value from visualization.