Papers.

Compromises in Healthcare Privacy due to Data Breaches

Healthcare privacy is essential for people because any leaked information could be used against the interests of the person in providing healthcare. The paper analyzes major healthcare data breaches globally, presents technology-based solutions to prevent such breaches, and concludes with policy guidelines for holders of health data to provide adequate data protection.

CFA: A Practical Prediction System for Video QoE Optimization

Many prior efforts have suggested that Internet video Quality of Experience (QoE) could be dramatically improved by using data-driven prediction of video quality for different choices (e.g., CDN or bitrate) to make optimal decisions. However, building such a prediction system is challenging on two fronts. First, the relationships between video quality and observed session features can be quite complex. Second, video quality changes dynamically. Thus, we need a prediction model that is (a) expressive enough to capture these complex relationships and (b) capable of updating quality predictions in near real-time. Unfortunately, several seemingly natural solutions (e.g., simple machine learning approaches and simple network models) fail on one or more fronts. Thus, the potential benefits promised by these prior efforts remain unrealized. We address these challenges and present the design and implementation of Critical Feature Analytics (CFA). The design of CFA is driven by domain-specific insights that video quality is typically determined by a small subset of critical features whose criticality persists over several tens of minutes. This enables a scalable and accurate workflow where we automatically learn critical features for different sessions on coarse-grained timescales, while updating quality predictions in near real-time. Using a combination of a real-world pilot deployment and trace-driven analysis, we demonstrate that CFA leads to significant improvements in video quality; e.g., 32% less buffering time and 12% higher bitrate than a random decision maker.

BlowFish: Dynamic Storage-Performance Tradeoff in Data Stores

Architecture-Led Safety Process

Architecture-Led Safety Analysis (ALSA) is a safety analysis method that uses early architecture knowledge to supplement traditional safety analysis techniques to identify faults as early as possible. The method begins by creating a definition of the operational environment within which the system under design will operate. ALSA uses the early architecture knowledge of the system and standardized error guide words to identify hazards in the system. These hazards are analyzed using knowledge of the architecture and safety requirements, intended to mitigate the hazards, that are added to the system’s requirements. ALSA continues its analysis down the full depth of the system implementation hierarchy. As additional implementation details are defined, the hazard analysis is applied to the subcomponents. ALSA also cuts across many of the phases in the development lifecycle. The hazard analysis feeds the requirements definition, architecture definition, and verification and validation phases.

Architecture-led Diagnosis and Verification of a Stepper Motor Controller

This paper discussed an architecture-led approach to diagnosing time sensitive issues with a stepper motor controller that manages fuel flow of an engine. A real engine control system design had originally been modeled and verified with SCADE . The potential for missed steps that result in misalignment in the fuel valve position is difficult to test for and was not discovered until after the engine went into operation. We utilize the execution and communication timing semantics of AADL to architecturally characterize the interaction between the elements of the stepper motor control systems. We then characterize the functional behavior in the context of the task dispatch and input handling semantics using the AADL Behavior Annex and identify potential fault sources and their impact using the AADL Error Model Annex. The identified the potential error sources, early arrival and mismatched command rates, we quantify the condition for this to occur and analyze the system based on timing data from scheduling analysis and actual timing measurements. We use this analysis to evaluate several proposed design corrections.

Architecture Fault Modeling and Analysis with the Error Model Annex, Version 2

Safety-critical software-reliant systems must manage component failures and conditions of anomalous interaction among components as hazards that affect a system's safety, reliability, and security so the potential effects of hazards on system operation are reduced to an acceptable risk. Standards and recommended practices for safety-critical systems outline methods for analysis, but security-related practices are typically addressed through separate guidance. This report provides guidance on using the Error Model Annex, Version 2 (EMV2), notation for architecture fault modeling and analysis, which supports automated safety, reliability, and security analyses from the same annotated architecture model to ensure consistency across analysis results. EMV2 augments architecture models expressed in the Architecture Analysis & Design Language with fault information to characterize anomalous conditions. The report introduces concepts for architecture fault modeling of systems in an operational environment at three levels of abstraction. In addition, EMV2 introduces the concept of error types to characterize exceptional conditions and their propagation. Finally, EMV2 allows users to specify which system components are expected to detect, report, and manage anomalous conditions and their propagation and to reflect the effects of recovery and repair actions as error behavior states. The report includes several example models.

Apache Spark: Scala or Python or Java, Oh My!

A Wardrobe for the Emperor: Stitching Practical Bias into Systems Software Research

A TASTY Alternative

A Requirement Specification Language for AADL

This report describes a textual requirement specification language, called ReqSpec, for the Architecture Analysis & Design Language (AADL). It is based on the draft Requirements Definition and Analysis Language Annex, which defines a meta-model for requirement specification as an-notations to AADL models. A set of plug-ins to the Open Source AADL Tool Environment (OSATE) toolset supports the ReqSpec language. Users can follow an architecture-led requirement specification process that uses AADL models to represent the system in its operational con-text as well as the architecture of the system of interest. ReqSpec can also be used to represent existing stakeholder and system requirement documents. Requirement documents represented in the Requirements Interchange Format can be imported into OSATE to migrate such documents into an architecture-centric virtual integration process. Finally, ReqSpec is an element of an architecture-led, incremental approach to system assurance. In this approach, requirements specifications are complemented with verification plans. When executed, these plans produce evidence that a system implementation satisfies the requirements. This report introduces the ReqSpec notation and illustrates its use on an example.

A Non-generative Framework and Convex Relaxations for Unsupervised Learning.

exploreAT! Perspektiven einer Transformation am Beispiel eines lexikographischen Jahrhundertprojekts

Modeling Genre With The Music Genome Project: Comparing Human-Labeled Attributes And Audio Features.

Improving Online Marketing Experiments with Drifting Multi-armed Bandits.

Automatic Abstraction of Flow of Control in a System of Distributed Software Components

CloudFoundry (CF) provides an open source platform-as-a-service software for deploying scalable software systems to the cloud. The architecture for CF is distributed by design and consists of several components which interact with one another through a message-oriented middleware. This message-oriented distributed design delivers on the scalability and resiliency requirements of the platform. In such a complex distributed multicomponent system, there is a steep learning curve for software developers to understand how components interact, what messages are exchanged between them, and how the message exchanges affect the behaviour of the system. In particular developers find it difficult to identify the execution flows, the authentication flows, interactions with the persistence layer, etc. We have developed a framework that allows interpreting the behaviour of the system by analyzing the exchanged messages between components, inspecting message contents, and extracting data and control flow across components. The paramount aim is to improve developers’ understandability of the system and to examine software resiliency through approaches like bug injection and message alterations. An initial version of our framework was released to the CF community and we have collected feedback that indeed show that we are achieving some of our goals.

TIDAQL - A Query Language Enabling on-Line Analytical Processing of Time Interval Data

Compile-Time Type-Driven Data Representation Transformations in Object-Oriented Languages

Language Support for Distributed Functional Programming

Decomposing Overcomplete 3rd Order Tensors using Sum-of-Squares Algorithms.

Distributed Programming via Safe Closure Passing

Part of Speech Annotation of Intermediate Versions in the Keystroke Logged Translation Corpus

Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks

Because of their superior ability to preserve sequence information over time, Long Short-Term Memory (LSTM) networks, a type of recurrent neural network with a more complex computational unit, have obtained strong results on a variety of sequence modeling tasks. The only underlying LSTM structure that has been explored so far is a linear chain. However, natural language exhibits syntactic properties that would naturally combine words to phrases. We introduce the Tree-LSTM, a generalization of LSTMs to tree-structured network topologies. Tree-LSTMs outperform all existing systems and strong LSTM baselines on two tasks: predicting the semantic relatedness of two sentences (SemEval 2014, Task 1) and sentiment classification (Stanford Sentiment Treebank).

WOLFE: An NLP-friendly Declarative Machine Learning Stack

Developing machine learning algorithms for natural language processing (NLP) applications is inherently an iterative process, involving a continuous refinement of the choice of model, engineering of features, selection of inference algorithms, search for the right hyperparameters, and error analysis. Existing probabilistic program languages (PPLs) only provide partial solutions; most of them do not support commonly used models such as matrix factorization or neural networks, and do not facilitate interactive and iterative programming that is crucial for rapid development of these models. In this demo we introduce WOLFE, a stack designed to facilitate the development of NLP applications: (1) the WOLFE language allows the user to concisely define complex models, enabling easy modification and extension, (2) the WOLFE interpreter transforms declarative machine learning code into automatically differentiable terms or, where applicable, into factor graphs that allow for complex models to be applied to real-world applications, and (3) the WOLFE IDE provides a number of different visual and interactive elements, allowing intuitive exploration and editing of the data representations, the underlying graphical models, and the execution of the inference algorithms.

CASSA: A Context-Aware Synonym Simplification Algorithm

Distributed Representations of Words to Guide Bootstrapped Entity Classifiers

Bootstrapped classifiers iteratively generalize from a few seed examples or prototypes to other examples of target labels. However, sparseness of language and limited supervision make the task difficult. We address this problem by using distributed vector representations of words to aid the generalization. We use the word vectors to expand entity sets used for training classifiers in a bootstrapped pattern-based entity extraction system. Our experiments show that the classifiers trained with the expanded sets perform better on entity extraction from four online forums, with 30% F1 improvement on one forum. The results suggest that distributed representations can provide good directions for generalization in a bootstrapping system.

Injecting Logical Background Knowledge into Embeddings for Relation Extraction

Matrix factorization approaches to relation extraction provide several attractive features: they support distant supervision, handle open schemas, and leverage unlabeled data. Unfortunately, these methods share a shortcoming with all other distantly supervised approaches: they cannot learn to extract target relations without existing data in the knowledge base, and likewise, these models are inaccurate for relations with sparse data. Rule-based extractors, on the other hand, can be easily extended to novel relations and improved for existing but inaccurate relations, through first-order formulae that capture auxiliary domain knowledge. However, usually a large set of such formulae is necessary to achieve generalization. In this paper, we introduce a paradigm for learning low-dimensional embeddings of entity-pairs and relations that combine the advantages of matrix factorization with first-order logic domain knowledge. We introduce simple approaches for estimating such embeddings, as well as a novel training algorithm to jointly optimize over factual and first-order logic information. Our results show that this method is able to learn accurate extractors with little or no distant supervision alignments, while at the same time generalizing to textual patterns that do not appear in the formulae.

Privacy Protection and Data Breaches

Scalability Challenges in Web Search Engines

Investigating consonant reduction in Mandarin Chinese with improved forced alignment

The effect of spectral slope on pitch perception

The Data Engineer

From DBpedia and WordNet hierarchies to LinkedIn and Twitter

Sentence selection for automatic scoring of Mandarin proficiency

Visually-Verifiable Textual Entailment: A Challenge Task for Combining Language and Vision

Kafka, Samza and the Unix Philosophy of Distributed Data

Catalyzing social media scholarship with open tools and data

Research Priorities for Robust and Beneficial Artificial Intelligence

Success in the quest for artificial intelligence has the potential to bring unprecedented benefits to humanity, and it is therefore worthwhile to investigate how to maximize these benefits while avoiding potential pitfalls. This article gives numerous examples (which should by no means be construed as an exhaustive list) of such worthwhile research aimed at ensuring that AI remains robust and beneficial.

Never-Ending Learning

Fictive Motion

Real Time Analytics: Algorithms and Systems

Velocity is one of the 4 Vs commonly used to characterize Big Data [5]. In this regard, Forrester remarked the following in Q3 2014 [8]: "The high velocity, white-water flow of data from innumerable real-time data sources such as market data, Internet of Things, mobile, sensors, click-stream, and even transactions remain largely unnavigated by most firms. The opportunity to leverage streaming analytics has never been greater." Example use cases of streaming analytics include, but not limited to: (a) visualization of business metrics in real-time (b) facilitating highly personalized experiences (c) expediting response during emergencies. Streaming analytics is extensively used in a wide variety of domains such as healthcare, e-commerce, financial services, telecommunications, energy and utilities, manufacturing, government and transportation. In this tutorial, we shall present an in-depth overview of streaming analytics -- applications, algorithms and platforms -- landscape. We shall walk through how the field has evolved over the last decade and then discuss the current challenges -- the impact of the other three Vs, viz., Volume, Variety and Veracity, on Big Data streaming analytics. The tutorial is intended for both researchers and practitioners in the industry. We shall also present state-of-the-affairs of streaming analytics at Twitter.

SAASFEE: Scalable Scientific Workflow Execution Engine

Scaling Spark in the Real World: Performance and Usability

Apache Spark is one of the most widely used open source processing engines for big data, with rich language-integrated APIs and a wide range of libraries. Over the past two years, our group has worked to deploy Spark to a wide range of organizations through consulting relationships as well as our hosted service, Databricks. We describe the main challenges and requirements that appeared in taking Spark to a wide set of users, and usability and performance improvements we have made to the engine in response.

Building a Replicated Logging System with Apache Kafka

An Enriched Ciphering Method to Evaluate Performance of EEA2-algorithm for LTE Security

Ancestry-constrained phylogenetic analysis supports the Indo-European steppe hypothesis

Identifying and Cultivating Superforecasters as a Method of Improving Probabilistic Predictions

Across a wide range of tasks, research has shown that people make poor probabilistic predictions of future events. Recently, the U.S. Intelligence Community sponsored a series of forecasting tournaments designed to explore the best strategies for generating accurate subjective probability estimates of geopolitical events. In this article, we describe the winning strategy: culling off top performers each year and assigning them into elite teams of superforecasters. Defying expectations of regression toward the mean 2 years in a row, superforecasters maintained high accuracy across hundreds of questions and a wide array of topics. We find support for four mutually reinforcing explanations of superforecaster performance: (a) cognitive abilities and styles, (b) task-specific skills, (c) motivation and commitment, and (d) enriched environments. These findings suggest that superforecasters are partly discovered and partly created-and that the high-performance incentives of tournaments highlight aspects of human judgment that would not come to light in laboratory paradigms focused on typical performance.

Psychological Language on Twitter Predicts County-Level Heart Disease Mortality

Hostility and chronic stress are known risk factors for heart disease, but they are costly to assess on a large scale. We used language expressed on Twitter to characterize community-level psychological correlates of age-adjusted mortality from atherosclerotic heart disease (AHD). Language patterns reflecting negative social relationships, disengagement, and negative emotions-especially anger-emerged as risk factors; positive emotions and psychological engagement emerged as protective factors. Most correlations remained significant after controlling for income and education. A cross-sectional regression model based only on Twitter language predicted AHD mortality significantly better than did a model that combined 10 common demographic, socioeconomic, and health risk factors, including smoking, diabetes, hypertension, and obesity. Capturing community psychological characteristics through social media is feasible, and these characteristics are strong markers of cardiovascular mortality at the community level.

Generalized reciprocal collision avoidance

Reciprocal collision avoidance has become a popular area of research over recent years. Approaches have been developed for a variety of dynamic systems ranging from single integrators to car-like, differential-drive, and arbitrary, linear equations of motion. In this paper, we present two contributions. First, we provide a unification of these previous approaches under a single, generalized representation using control obstacles. In particular, we show how velocity obstacles, acceleration velocity obstacles, continuous control obstacles, and LQR-obstacles are special instances of our generalized framework. Secondly, we present an extension of control obstacles to general reciprocal collision avoidance for non-linear, non-homogeneous systems where the robots may have different state spaces and different non-linear equations of motion from one another. Previous approaches to reciprocal collision avoidance could not be applied to such systems, as they use a relative formulation of the equations of motion and can, therefore, only apply to homogeneous, linear systems where all robots have the same linear equations of motion. Our approach allows for general mobile robots to independently select new control inputs while avoiding collisions with each other. We implemented our approach in simulation for a variety of mobile robots with non-linear equations of motion: differential-drive, differential-drive with a trailer, car-like, and hovercrafts. We also performed physical experiments with a combination of differential-drive, differential-drive with a trailer, and car-like robots. Our results show that our approach is capable of letting a non-homogeneous group of robots with non-linear equations of motion safely avoid collisions at real-time computation rates.

Design Challenges for Entity Linking

Recent research on entity linking (EL) has introduced a plethora of promising techniques, ranging from deep neural networks to joint inference. But despite numerous papers there is surprisingly little understanding of the state of the art in EL. We attack this confusion by analyzing differences between several versions of the EL problem and presenting a simple yet effective, modular, unsupervised system, called Vinculum, for entity linking. We conduct an extensive evaluation on nine data sets, comparing Vinculum with two state-of-the-art systems, and elucidate key aspects of the system that include mention extraction, candidate generation, entity type prediction, entity coreference, and coherence.

Learning a Compositional Semantics for Freebase with an Open Predicate Vocabulary