Papers.

Lexical Quality as a Measure for Textual Web Accessibility

Theory Presentation Combinators

Usage Data in Web Search: Benefits and Limitations

Pause 'n' Play: Formalizing Asynchronous C#

Reducing the History in Decentralized Interaction-Based Reputation Systems

Adaptive Information Integration: Bridging the Semantic Gap between Numerical Simulations

Modeling Static Caching in Web Search Engines

Advances in information retrieval : 34th European Conference on IR Research, ECIR 2012, Barcelona, Spain, April 1-5 2012 : proceedings

Linking Localisation and Language Resources

Terminological Systems

Implementations of Terminology

History of Terminology and Terminological Logics

Conclusion

Introduction

Design, implementation and verification of MILS systems

SUMMARY Safety‐critical systems are used in many domains (military, avionics, aerospace, etc.) and handle critical data in hostile environments. To prevent data access by unauthorized subjects, they must protect and isolate information so that only allowed entities can read or write information. However, because of their increased number of functionalities, safety‐critical systems design becomes more complex; this increases difficulties in the design and the verification of security functions and potential error in their implementation. The multiple independent levels of security (MILS) approach introduces rules and guidelines for the design of secure systems. It isolates data according to their security levels, reducing system complexity to ease development. However, there is no approach addressing the whole development of MILS systems from high‐level specification (application components with their security levels) to the final implementation (code that executes application functions and provide security mechanisms). This paper presents a complete development approach for the design, verification and implementation of MILS architectures. It aims at providing a complete framework to build secure applications based on MILS guidelines. We describe security concerns using a modeling language, verify security requirements and automatically implement the system code generation techniques and a MILS‐compliant operating system that provides security functions. Copyright © 2012 John Wiley & Sons, Ltd.

Truthy: Enabling the Study of Online Social Networks

The End of an Architectural Era for Analytical Databases

Traditional enterprise warehouse solutions center around an analytical database system that is monolithic and inflexible: data needs to be extracted, transformed, and loaded into the rigid relational form before analysis. It takes years of sophisticated planning to provision and deploy a warehouse; adding new hardware resources to an existing warehouse is an equally lengthy and daunting task. Additionally, modern data analysis employs statistical methods that go well beyond the typical roll-up and drill-down capabilities provided by warehouse systems. Although it is possible to implement such methods using a combination of SQL and UDFs, query engines in relational databases are ill-suited for these. The Hadoop ecosystem introduces a suite of tools for data analytics that overcome some of the problems of traditional solutions. These systems, however, forgo years of warehouse research. Memory is significantly underutilized in Hadoop clusters, and execution engine is naive compared with its relational counterparts. It is time to rethink the design of data warehouse systems and take the best from both worlds. The new generation of warehouse systems should be modular, high performance, fault-tolerant, easy to provision, and designed to support both SQL query processing and machine learning applications. This paper references the Shark system developed at Berkeley as an initial attempt.

Opening the Black Boxes in Data Flow Optimization

A New Theory of Anyons

Probabilistically Bounded Staleness for Practical Partial Quorums

Learning to Rank Query Recommendations by Semantic Similarities

Scene Parsing with Multiscale Feature Learning, Purity Trees, and Optimal Covers

Scene parsing, or semantic segmentation, consists in labeling each pixel in an image with the category of the object it belongs to. It is a challenging task that involves the simultaneous detection, segmentation and recognition of all the objects in the image. The scene parsing method proposed here starts by computing a tree of segments from a graph of pixel dissimilarities. Simultaneously, a set of dense feature vectors is computed which encodes regions of multiple sizes centered on each pixel. The feature extractor is a multiscale convolutional network trained from raw pixels. The feature vectors associated with the segments covered by each node in the tree are aggregated and fed to a classifier which produces an estimate of the distribution of object categories contained in the segment. A subset of tree nodes that cover the image are then selected so as to maximize the average "purity" of the class distributions, hence maximizing the overall likelihood that each segment will contain a single object. The convolutional network feature extractor is trained end-to-end from raw pixels, alleviating the need for engineered features. After training, the system is parameter free. The system yields record accuracies on the Stanford Background Dataset (8 classes), the Sift Flow Dataset (33 classes) and the Barcelona Dataset (170 classes) while being an order of magnitude faster than competing approaches, producing a 320 \times 240 image labeling in less than 1 second.

Visualizing Communication on Social Media: Making Big Data Accessible

WizIE: A Best Practices Guided Development Environment for Information Extraction

Web Services Data Access and Integration – The Relational Realisation (WS-DAIR) Specification

Web Services Data Access and Integration – The Core (WS-DAI) Specification

Weakly Supervised Training of Semantic Parsers

Validate implementation correctness using simulation: the TASTE approach

Abstract: High-integrity systems operate in hostile environment and must guarantee a continuous operational state, even if unexpected events happen. In addition, these systems have stringent requirements that must be validated and correctly translated from high-level specifications down to code. All these constraints make the overall development process more time-consuming. This becomes especially complex because the number of system functions keeps increasing over the years. As a result, engineers must validate system implementation and check that its execution conforms to the specifications. To do so, a traditional approach consists in a manual instrumentation of the implementation code to trace system activity while operating. However, this might be error-prone because modifications are not automatic and still made manually. Furthermore, such modifications may have an impact on the actual behavior of the system. In this paper, we present an approach to validate a system implementation by comparing execution against simulation. In that purpose, we adapt TASTE, a set of tools that eases system development by automating each step as much as possible. In particular, TASTE automates system implementation from functional (system functions description with their properties – period, deadline, priority, etc.) and deployment (processors, buses, devices to be used) models. We tailored this tool-chain to create traces during system execution. Generated output shows activation time of each task, usage of communication ports (size of the queues, instant of events pushed/pulled, etc.) and other relevant execution metrics to be monitored. As a consequence, system engineers can check implementation correctness by comparing simulation and execution metrics.

Transnational Connections| New ICTs and the Study of Political Communication

Training Factored PCFGs with Expectation Propagation

Tracking Epidemics with Natural Language Processing and Crowdsourcing

Towards Short-Lived Certificates

TinyToCS Volume 1 Chairs' Note.

The Dynamic Relationship between Entrepreneurship , Unemployment , and Growth : Evidence from U . S . Industries

The 2nd Temporal Web Analytics Workshop (TempWeb)

TASTE: An open-source tool-chain for embedded system and software development

This paper presents the results of the past two years of continuous development of the TASTE tool-set. TASTE is a development environment dedicated to embedded, real-time systems and was created under the initiative of the European Space Agency back in 2008, after the completion of a FP6 project called ASSERT. TASTE is free and open source, and is currently used to design small to medium-size systems; it relies on two powerful and complementary modeling languages: ASN.1 and AADL, and comes together with a solid engineering approach. TASTE brings ideas on how a system can be optimally built, by taking and putting together components of heterogeneous nature, and making sure that they run according to their specification and without software hacking introduced during the development. The latest TASTE improvements include several major features such as the integration of VHDL components, the recording/replay of runtime scenarii using message sequence charts, the support of legacy encoding protocols, means to inspect and patch data at runtime, and the generation of systems for additional platforms such as RT Linux (Xenomai). Our graphical front-ends have also been redesigned to offer a better user experience.

Speaking with the crowd

Serving large-scale batch computed data with project Voldemort

Semantic Models for Question Answering.

Abstract. The research presented in this paper focuses on the adoption of semantic models for Question Answering (QA) systems. We propose a framework which exploits semantic technologies to analyze the question, retrieve and rank relevant passages. It exploits: (a) Natural Language Processing algorithmsfortheanalysisofquestionsandcandidateanswers both in English and Italian; (b) Information Retrieval (IR) probabilistic models for retrieving candidate answers and (c) Machine Learning methods for question classification. The data source for the answers is an unstructured text document collection stored in search indices. The aim of the research is to improve the system performances by introducing semantic models in every step of the answering process. 1

Semantic Compositionality through Recursive Matrix-Vector Spaces

Search for Excited Randall-Sundrum Gravitons from Warped Extra Dimensions with Semi-Leptonic Diboson Final States using the ATLAS detector at the LHC

Scala Macros, a Technical Report

Scala in Depth

Sample Complexity of Composite Likelihood

We present the first PAC bounds for learning parameters of Conditional Random Fields [12] with general structures over discrete and real-valued variables. Our bounds apply to composite likelihood [14], which generalizes maximum likelihood and pseudolikelihood [3]. Moreover, we show that the only existing algorithm with a PAC bound for learning high-treewidth discrete models [1] can be viewed as a computationally inefficient method for computing pseudolikelihood. We present an extensive empirical study of the statistical efficiency of these estimators, as predicted by our bounds. Finally, we use our bounds to show how to construct computationally and statistically efficient composite likelihood estimators.

Resilient Distributed Datasets: A Fault-Tolerant Abstraction for In-Memory Cluster Computing

We present Resilient Distributed Datasets (RDDs), a distributed memory abstraction that lets programmers perform in-memory computations on large clusters in a fault-tolerant manner. RDDs are motivated by two types of applications that current computing frameworks handle inefficiently: iterative algorithms and interactive data mining tools. In both cases, keeping data in memory can improve performance by an order of magnitude. To achieve fault tolerance efficiently, RDDs provide a restricted form of shared memory, based on coarsegrained transformations rather than fine-grained updates to shared state. However, we show that RDDs are expressive enough to capture a wide class of computations, including recent specialized programming models for iterative jobs, such as Pregel, and new applications that these models do not capture. We have implemented RDDs in a system called Spark, which we evaluate through a variety of user applications and benchmarks. 1

QuestionCube: a Framework for Question Answering.

QuestionCube is a framework for Question Answering (QA) that combines several techniques to retrieve passages containing the exact answers for natural language questions. It exploits: (a) Natural Language Processing algorithms for question and candidate answers analysis both in English and Italian; (b) Information Retrieval probabilistic models for candidate answers retrieval and (c) Machine Learning methods for question classification. The data source for the answer is an unstructured text document collection stored in search indices. In this paper an overview of the QuestionCube framework architecture is provided, together with a description of Wikiedi, a QA system for Wikipedia which exploits the proposed framework.

Python Pocket Primer

Project Kepler: Compile-Time Metaprogramming for Scala

Programming Hive

Proceedings of the 34th European conference on Advances in Information Retrieval