A constrained $$\ell $$1 minimization approach for estimating multiple sparse Gaussian or nonparanormal graphical models

Identifying context-specific entity networks from aggregated data is an important task, arising often in bioinformatics and neuroimaging applications. Computationally, this task can be formulated as jointly estimating multiple different, but related, sparse undirected graphical models (UGM) from aggregated samples across several contexts. Previous joint-UGM studies have mostly focused on sparse Gaussian graphical models (sGGMs) and can’t identify context-specific edge patterns directly. We, therefore, propose a novel approach, SIMULE (detecting Shared and Individual parts of MULtiple graphs Explicitly) to learn multi-UGM via a constrained \(\ell \)1 minimization. SIMULE automatically infers both specific edge patterns that are unique to each context and shared interactions preserved among all the contexts. Through the \(\ell \)1 constrained formulation, this problem is cast as multiple independent subtasks of linear programming that can be solved efficiently in parallel. In addition to Gaussian data, SIMULE can also handle multivariate Nonparanormal data that greatly relaxes the normality assumption that many real-world applications do not follow. We provide a novel theoretical proof showing that SIMULE achieves a consistent result at the rate \(O(\log (Kp)/n_{tot})\). On multiple synthetic datasets and two biomedical datasets, SIMULE shows significant improvement over state-of-the-art multi-sGGM and single-UGM baselines (SIMULE implementation and the used datasets @https://github.com/QData/SIMULE).     

Fusion that matters: convolutional fusion networks for visual recognition

In recent years, deep learning has been successfully applied to diverse multimedia research areas, with the aim of learning powerful and informative representations for a variety of visual recognition tasks. In this work, we propose convolutional fusion networks (CFN) to integrate multi-level deep features and fuse a richer visual representation. Despite recent advances in deep fusion networks, they still have limitations due to expensive parameters and weak fusion modules. Instead, CFN uses 1 × 1 convolutional layers and global average pooling to generate side branches with few parameters, and employs a locally-connected fusion module, which can learn adaptive weights for different side branches and form a better fused feature. Specifically, we introduce three key components of the proposed CFN, and discuss its differences from other deep models. Moreover, we propose fully convolutional fusion networks (FCFN) that are an extension of CFN for pixel-level classification applied to several tasks, such as semantic segmentation and edge detection. Our experiments demonstrate that CFN (and FCFN) can achieve promising performance by consistent improvements for both image-level and pixel-level classification tasks, compared to a plain CNN. We release our codes on https://github.com/yuLiu24/CFN. Also, we make a live demo (goliath.liacs.nl) using a CFN model trained on the ImageNet dataset.     

Soft error resilience of Big Data kernels through algorithmic approaches

As the volume of data generated each day continues to increase, more and more interest is put into Big Data algorithms and the insight they provide.? Since these analyses require a substantial amount of resources, including physical machines, power, and time, reliable execution of the algorithms becomes critical. This paper analyzes the error resilience of a select group of popular Big Data algorithms and shows how they can effectively be made more fault-tolerant. Using KULFI (http://github.com/quadpixels/kulfi, 2013) and the LLVM (Proceedings of the 2004 international symposium on code generation and optimization (CGO 2004), San Jose, CA, USA, 2004) compiler for compilation allows injection of artificial soft faults throughout these algorithms, giving a thorough analysis of how faults in different locations can affect the outcome of the program. This information is then used to help guide incorporating fault tolerance mechanisms into the program, making them as impervious as possible to soft faults.     

Personalized multimedia content delivery on an interactive table by passive observation of museum visitors

The amount of multimedia data collected in museum databases is growing fast, while the capacity of museums to display information to visitors is acutely limited by physical space. Museums must seek the perfect balance of information given on individual pieces in order to provide sufficient information to aid visitor understanding while maintaining sparse usage of the walls and guaranteeing high appreciation of the exhibit. Moreover, museums often target the interests of average visitors instead of the entire spectrum of different interests each individual visitor might have. Finally, visiting a museum should not be an experience contained in the physical space of the museum but a door opened onto a broader context of related artworks, authors, artistic trends, etc. In this paper we describe the MNEMOSYNE system that attempts to address these issues through a new multimedia museum experience. Based on passive observation, the system builds a profile of the artworks of interest for each visitor. These profiles of interest are then used to drive an interactive table that personalizes multimedia content delivery. The natural user interface on the interactive table uses the visitor’s profile, an ontology of museum content and a recommendation system to personalize exploration of multimedia content. At the end of their visit, the visitor can take home a personalized summary of their visit on a custom mobile application. In this article we describe in detail each component of our approach as well as the first field trials of our prototype system built and deployed at our permanent exhibition space at LeMurate (http://www.lemurate.comune.fi.it/lemurate/) in Florence together with the first results of the evaluation process during the official installation in the National Museum of Bargello (http://www.uffizi.firenze.it/musei/?m=bargello).     

Modulo Based Data Placement Algorithm for Energy Consumption Optimization of MapReduce System

With the explosion of data production, the efficiency of data management and analysis has been concerned by both industry and academia. Meanwhile, more and more energy is consumed by the IT infrastructure especially the larger scale distributed systems. In this paper, a novel idea for optimizing the Energy Consumption (EC for short) of MapReduce system is proposed. We argue that a fair data placement is helpful to save energy, and then we propose three goals of data placement, and a modulo based Data Placement Algorithm (DPA for short) which achieves these goals. Afterwards, the correctness of the proposed DPA is proved from both theoretical and experimental perspectives. Three different systems which implement MapReduce model with different DPAs are compared in our experiments. Our algorithm is proved to optimize EC effectively, without introducing the additional costs and delaying data loading. With the help of our DPA, the EC for the WordCount (https://src/examples/org/apache/hadoop/examples/), Sort (https://src/examples/org/apache/hadoop/examples/sort) and MRBench (https://src/examples/org/apache/hadoop/mapred/) can be reduced by 10.9 %, 8.3 % and 17 % respectively, and time consumption is reduced by 7 %, 6.3 % and 7 % respectively.     

Benchmarking SQL on MapReduce systems using large astronomy databases

In the era of bigdata, with a massive set of digital information of unprecedented volumes being collected and/or produced in several application domains, it becomes more and more difficult to manage and query large data repositories. In the framework of the PetaSky project (http://com.isima.fr/Petasky), we focus on the problem of managing scientific data in the field of cosmology. The data we consider are those of the LSST project (http://www.lsst.org/). The overall size of the database that will be produced is expected to exceed 60 PB (Lsst data challenge handbook, 2012). In order to evaluate the performances of existing SQL On MapReduce data management systems, we conducted extensive experiments by using data and queries from the area of cosmology. The goal of this work is to report on the ability of such systems to support large scale declarative queries. We mainly investigated the impact of data partitioning, indexing and compression on query execution performances.     

Improving topology optimization intuition through games

This paper describes the educational game, TopOpt Game, which invites the player to solve various optimization challenges. The main purpose of gamifying topology optimization is to create a supplemental educational tool which can be used to introduce concepts of topology optimization to newcomers as well as to train human intuition of topology optimization. The players are challenged to solve the standard minimum compliance problem in 2D by distributing material in a design domain given a number of loads and supports with a material constraint. A statistical analysis of the gameplay data shows that players achieve higher scores the more they play the game. The game is freely available for the iOS platform at Apple’s App Store and at http://www.topopt.dtu.dk/?q=node/909 for Windows and OSX.     

VQA: Visual Question Answering

We propose the task of free-form and open-ended Visual Question Answering (VQA). Given an image and a natural language question about the image, the task is to provide an accurate natural language answer. Mirroring real-world scenarios, such as helping the visually impaired, both the questions and answers are open-ended. Visual questions selectively target different areas of an image, including background details and underlying context. As a result, a system that succeeds at VQA typically needs a more detailed understanding of the image and complex reasoning than a system producing generic image captions. Moreover, VQA is amenable to automatic evaluation, since many open-ended answers contain only a few words or a closed set of answers that can be provided in a multiple-choice format. We provide a dataset containing \(\sim \)0.25 M images, \(\sim \)0.76 M questions, and \(\sim \)10 M answers (www.visualqa.org), and discuss the information it provides. Numerous baselines and methods for VQA are provided and compared with human performance. Our VQA demo is available on CloudCV (http://cloudcv.org/vqa).     

Cheap talk,cooperation, and trust in global software engineering

In our studies of global software engineering (GSE) teams, we found that informal, non-work-related conversations are positively associated with trust. Seeking to use novel analytical techniques to more carefully investigate this phenomenon, we described these non-work-related conversations by adapting the economics literature concept of “cheap talk,” and studied it using Evolutionary Game Theory (EGT). More specifically, we modified the classic Stag-hunt game and analyzed the dynamics in a fixed population setting (an abstraction of a GSE team). Doing so, we were able to demonstrate how cheap talk over the Internet (e-cheap talk) was powerful enough to facilitate the emergence of trust and improve the probability of cooperation where the punishment for uncooperative behavior is comparable to the cost of the cheap talk. To validate the results of our theoretical approach, we conducted two empirical case studies that analyzed the logged IRC development discussions of Apache Lucene (http://lucene.apache. org/) and Chromium OS (http://www.chromium.org/chromium-os) using both quantitative and qualitative methods. The results provide general support to the theoretical propositions. We discuss our findings and the theoretical and practical implications to GSE collaborations and research.     

An Embedding Based IR Model for Disaster Situations

Twitter (http://twitter.com) is one of the most popular social networking platforms. Twitter users can easily broadcast disaster-specific information, which, if effectively mined, can assist in relief operations. However, the brevity and informal nature of tweets pose a challenge to Information Retrieval (IR) researchers. In this paper, we successfully use word embedding techniques to improve ranking for ad-hoc queries on microblog data. Our experiments with the ‘Social Media for Emergency Relief and Preparedness’ (SMERP) dataset provided at an ECIR 2017 workshop show that these techniques outperform conventional term-matching based IR models. In addition, we show that, for the SMERP task, our word embedding based method is more effective if the embeddings are generated from the disaster specific SMERP data, than when they are trained on the large social media collection provided for the TREC (http://trec.nist.gov/) 2011 Microblog track dataset.     

Pointwise and pairwise clothing annotation: combining features from social media

In this paper, we present effective algorithms to automatically annotate clothes from social media data, such as Facebook and Instagram. Clothing annotation can be informally stated as recognizing, as accurately as possible, the garment items appearing in the query photo. This task brings huge opportunities for recommender and e-commerce systems, such as capturing new fashion trends based on which clothes have been used more recently. It also poses interesting challenges for existing vision and recognition algorithms, such as distinguishing between similar but different types of clothes or identifying a pattern of a cloth even if it has different colors and shapes. We formulate the annotation task as a multi-label and multi-modal classification problem: (i) both image and textual content (i.e., tags about the image) are available for learning classifiers, (ii) the classifiers must recognize a set of labels (i.e., a set of garment items), and (iii) the decision on which labels to assign to the query photo comes from a set of instances that is used to build a function, which separates labels that should be assigned to the query photo, from those that should not be assigned. Using this configuration, we propose two approaches: (i) the pointwise one, called MMCA, which receives a single image as input, and (ii) a multi-instance classification, called M3CA, also known as pairwise approach, which uses pair of images to create the classifiers. We conducted a systematic evaluation of the proposed algorithms using everyday photos collected from two major fashion-related social media, namely pose.com and chictopia.com. Our results show that the proposed approaches provide improvements when compared to popular first choice multi-label, multi-modal, multi-instance algorithms that range from 20 % to 30 % in terms of accuracy.     

Aspect identification and ratings inference for hotel reviews

Today, a large volume of hotel reviews is available on many websites, such as TripAdvisor (http://www.tripadvisor.com) and Orbitz (http://www.orbitz.com). A typical review contains an overall rating, several aspect ratings, and review text. The rating is an abstract of review in terms of numerical points. The task of aspect-based opinion summarization is to extract aspect-specific opinions hidden in the reviews which do not have aspect ratings, so that users can quickly digest them without actually reading through them. The task consists of aspect identification and aspect rating inference. Most existing studies cannot utilize aspect ratings which become increasingly abundant on review hosts. In this paper, we propose two topic models which explicitly model aspect ratings as observed variables to improve the performance of aspect rating inference on unrated reviews. The experiment results show that our approaches outperform the existing methods on the data set crawled from TripAdvisor website.     

Estimating Dynamic Binary Panel Data Model with Random Effects: A Computational Note

Recently, Gao et al. (J Time Ser Anal, 2016 doi:  10.1111/jtsa.12178) propose a new estimation method for dynamic panel probit model with random effects, where the theoretical properties of estimator are derived. In this paper, we extend their estimation method to the \(T\ge 3\) case, and some Monte Carlo simulations are presented to illustrate the extended estimator.     

On pseudo-BL-algebras and pseudo-hoops with normal maximal filters

We study the class of pseudo-BL-algebras whose every maximal filter is normal. We present an equational base for this class and we extend these results for the class of basic pseudo hoops with fixed strong unit. This is a continuation of the research from Botur et al. (Soft Comput 16:635–644, doi: 10.1007/s00500-011-0763-7, 2012).     

Dynamically pre-trained deep recurrent neural networks using environmental monitoring data for predicting PM<Subscript>2.5</Subscript>

Fine particulate matter (\(\hbox {PM}_{2.5}\)) has a considerable impact on human health, the environment and climate change. It is estimated that with better predictions, US$9 billion can be saved over a 10-year period in the USA (State of the science fact sheet air quality. http://www.noaa.gov/factsheets/new, 2012). Therefore, it is crucial to keep developing models and systems that can accurately predict the concentration of major air pollutants. In this paper, our target is to predict \(\hbox {PM}_{2.5}\) concentration in Japan using environmental monitoring data obtained from physical sensors with improved accuracy over the currently employed prediction models. To do so, we propose a deep recurrent neural network (DRNN) that is enhanced with a novel pre-training method using auto-encoder especially designed for time series prediction. Additionally, sensors selection is performed within DRNN without harming the accuracy of the predictions by taking advantage of the sparsity found in the network. The numerical experiments show that DRNN with our proposed pre-training method is superior than when using a canonical and a state-of-the-art auto-encoder training method when applied to time series prediction. The experiments confirm that when compared against the \(\hbox {PM}_{2.5}\) prediction system VENUS (National Institute for Environmental Studies. Visual Atmospheric Environment Utility System. http://envgis5.nies.go.jp/osenyosoku/, 2014), our technique improves the accuracy of \(\hbox {PM}_{2.5}\) concentration level predictions that are being reported in Japan.     

RST Signalling Corpus: a corpus of signals of coherence relations

We present the RST Signalling Corpus (Das et al. in RST signalling corpus, LDC2015T10. https://catalog.ldc.upenn.edu/LDC2015T10, 2015), a corpus annotated for signals of coherence relations. The corpus is developed over the RST Discourse Treebank (Carlson et al. in RST Discourse Treebank, LDC2002T07. https://catalog.ldc.upenn.edu/LDC2002T07, 2002) which is annotated for coherence relations. In the RST Signalling Corpus, these relations are further annotated with signalling information. The corpus includes annotation not only for discourse markers which are considered to be the most typical (or sometimes the only type of) signals in discourse, but also for a wide array of other signals such as reference, lexical, semantic, syntactic, graphical and genre features as potential indicators of coherence relations. We describe the research underlying the development of the corpus and the annotation process, and provide details of the corpus. We also present the results of an inter-annotator agreement study, illustrating the validity and reproducibility of the annotation. The corpus is available through the Linguistic Data Consortium, and can be used to investigate the psycholinguistic mechanisms behind the interpretation of relations through signalling, and also to develop discourse-specific computational systems such as discourse parsing applications.     

Multi-agent programming contest 2017

We present the twelfth edition of the Multi-Agent Programming Contest (https://multiagentcontest.org), an annual, community-serving competition that attracts groups from all over the world. Our contest facilitates comparison of multi-agent systems and provides a concrete problem that is interesting in itself and well-suited to be tackled in educational environments. This time, seven teams competed using strictly agent-based as well as traditional programming approaches.     

The Train Benchmark: cross-technology performance evaluation of continuous model queries

In model-driven development of safety-critical systems (like automotive, avionics or railways), well-formedness of models is repeatedly validated in order to detect design flaws as early as possible. In many industrial tools, validation rules are still often implemented by a large amount of imperative model traversal code which makes those rule implementations complicated and hard to maintain. Additionally, as models are rapidly increasing in size and complexity, efficient execution of validation rules is challenging for the currently available tools. Checking well-formedness constraints can be captured by declarative queries over graph models, while model update operations can be specified as model transformations. This paper presents a benchmark for systematically assessing the scalability of validating and revalidating well-formedness constraints over large graph models. The benchmark defines well-formedness validation scenarios in the railway domain: a metamodel, an instance model generator and a set of well-formedness constraints captured by queries, fault injection and repair operations (imitating the work of systems engineers by model transformations). The benchmark focuses on the performance of query evaluation, i.e. its execution time and memory consumption, with a particular emphasis on reevaluation. We demonstrate that the benchmark can be adopted to various technologies and query engines, including modeling tools; relational, graph and semantic databases. The Train Benchmark is available as an open-source project with continuous builds from https://github.com/FTSRG/trainbenchmark.     

Document segmentation and classification into musical scores and text

A new algorithm for segmenting documents into regions containing musical scores and text is proposed. Such segmentation is a required step prior to applying optical character recognition and optical music recognition on scanned pages that contain both music notation and text. Our segmentation technique is based on the bag-of-visual-words representation followed by random block voting (RBV) in order to detect the bounding boxes containing the musical score and text within a document image. The RBV procedure consists of extracting a fixed number of blocks whose position and size are sampled from a discrete uniform distribution that “over”-covers the input image. Each block is automatically classified as either coming from musical score or text and votes with a particular posterior probability of classification in its spatial domain. An initial coarse segmentation is obtained by summarizing all the votes in a single image. Subsequently, the final segmentation is obtained by subdividing the image in microblocks and classifying them using a N-nearest neighbor classifier which is trained using the coarse segmentation. We demonstrate the potential of the proposed method by experiments on two different datasets. One is on a challenging dataset of images collected and artificially combined and manipulated for this project. The other is a music dataset obtained by the scanning of two music books. The results are reported using precision/recall metrics of the overlapping area with respect to the ground truth. The proposed system achieves an overall averaged F-measure of 85 %. The complete source code package and associated data are available at https://github.com/fpeder/mscr under the FreeBSD license to support reproducibility.     

Evolutionary Frequency and Forecasting Accuracy: Simulations Based on an Agent-Based Artificial Stock Market

Learning how to forecast is always important for traders, and divergent learning frequencies prevail among traders. The influence of the evolutionary frequency on learning performance has occasioned many studies of agent-based computational finance (e.g., Lettau in J Econ Dyn Control 21:1117–1147, 1997. doi: 10.1016/S0165-1889(97)00046-8; Szpiro in Complexity 2(4):31–39, 1997. doi: 10.1002/(SICI)1099-0526(199703/04)2:4<31::AID-CPLX8>3.0.CO;2-3; Cacho and Simmons in Aust J Agric Resour Econ 43(3):305–322, 1999. doi: 10.1111/1467-8489.00081). Although these studies all suggest that evolving less frequently and, hence, experiencing more realizations help learning, this implication may result from their common stationary assumption. Therefore, we first attempt to approach this issue in a ‘dynamically’ evolving market in which agents learn to forecast endogenously generated asset prices. Moreover, in these studies’ market settings, evolving less frequently also meant having a longer time horizon. However, it is not true in many market settings that are even closer to the real financial markets. The clarification that the evolutionary frequency and the time horizon are two separate notions leaves the effect of the evolutionary frequency on learning even more elusive and worthy of exploration independently. We find that the influence of a trader’s evolutionary frequency on his forecasting accuracy depends on all market participants and the resulting price dynamics. In addition, prior studies also commonly assume that traders have identical preferences, which is too strong an assumption to apply to a real market. Considering the heterogeneity of preferences, we find that converging to the rational expectations equilibrium is hardly possible, and we even suggest that agents in a slow-learning market learn frequently. We also apply a series of econometric tests to explain the simulation results.