Suchergebnisse

With the advent of social networks and the rise of mobile technologies, users have become ubiquitous sensors capable of monitoring various real-world events in a crowd-sourced manner. Location-based social networks have proven to be faster than traditional media channels in reporting and geo-locating breaking news, i.e. Osama Bin Laden's death was first confirmed on Twitter even before the announcement from the communication department at the White House. However, the deluge of user-generated data on these networks requires intelligent systems capable of identifying and characterizing such events in a comprehensive manner. The data mining community coined the term, event detection , to refer to the task of uncovering emerging patterns in data streams . Nonetheless, most data mining techniques do not reproduce the underlying data generation process, hampering to self-adapt in fast-changing scenarios. Because of this, we propose a probabilistic machine learning approach to event detection which explicitly models the data generation process and enables reasoning about the discovered events. With the aim to set forth the differences between both approaches, we present two techniques for the problem of event detection in Twitter : a data mining technique called Tweet-SCAN and a machine learning technique called Warble. We assess and compare both techniques in a dataset of tweets geo-located in the city of Barcelona during its annual festivities. Last but not least, we present the algorithmic changes and data processing frameworks to scale up the proposed techniques to big data workloads. ; This work is partially supported by Obra Social "la Caixa", by the Spanish Ministry of Science and Innovation under contract (TIN2015-65316), by the Severo Ochoa Program (SEV2015-0493), by SGR programs of the Catalan Government (2014-SGR-1051, 2014-SGR-118), Collectiveware (TIN2015-66863-C2-1-R) and BSC/UPC NVIDIA GPU Center of Excellence.We would also like to thank the reviewers for their constructive feedback. ; Peer Reviewed ; Postprint (author's final draft)

With the advent of social networks and the rise of mobile technologies, users have become ubiquitous sensors capable of monitoring various real-world events in a crowd-sourced manner. Location-based social networks have proven to be faster than traditional media channels in reporting and geo-locating breaking news, i.e. Osama Bin Laden's death was first confirmed on Twitter even before the announcement from the communication department at the White House. However, the deluge of user-generated data on these networks requires intelligent systems capable of identifying and characterizing such events in a comprehensive manner. The data mining community coined the term, event detection , to refer to the task of uncovering emerging patterns in data streams . Nonetheless, most data mining techniques do not reproduce the underlying data generation process, hampering to self-adapt in fast-changing scenarios. Because of this, we propose a probabilistic machine learning approach to event detection which explicitly models the data generation process and enables reasoning about the discovered events. With the aim to set forth the differences between both approaches, we present two techniques for the problem of event detection in Twitter : a data mining technique called Tweet-SCAN and a machine learning technique called Warble. We assess and compare both techniques in a dataset of tweets geo-located in the city of Barcelona during its annual festivities. Last but not least, we present the algorithmic changes and data processing frameworks to scale up the proposed techniques to big data workloads. ; This work is partially supported by Obra Social "la Caixa", by the Spanish Ministry of Science and Innovation under contract (TIN2015-65316), by the Severo Ochoa Program (SEV2015-0493), by SGR programs of the Catalan Government (2014-SGR-1051, 2014-SGR-118), Collectiveware (TIN2015-66863-C2-1-R) and BSC/UPC NVIDIA GPU Center of Excellence.We would also like to thank the reviewers for their constructive feedback. ; Peer Reviewed ...

The geographical identification of content in Social Networks have enabled to bridge the gap between online social platforms and the physical world. Although vast amounts of data in such networks are due to breaking news or global occurrences, local events witnessed by users in situ are also present in these streams and of great importance for many city entities. Nowadays, unsupervised machine learning techniques, such as Tweet-SCAN, are able to retrospectively detect these local events from tweets. However, these approaches have limited abilities to reason about unseen observations in a principled way due to the lack of a proper probabilistic foundation. Probabilistic models have also been proposed for the task, but their event identification capabilities are far from those of Tweet-SCAN. In this paper, we identify two key factors which, when combined, boost the accuracy of such models. As a first key factor, we notice that the large amount of meaningless social data requires explicitly modeling non-event observations.Therefore, we propose to incorporate a background model that captures spatio-temporal fluctuations of non-event tweets. As a second key factor, we observe that the shortness of tweets hampers the application of traditional topic models. Thus, we integrate event detection and topic modeling, assigning topic proportions to events instead of assigning them to individual tweets. As a result, we propose Warble, a new probabilistic model and learning scheme for retrospective event detection that incorporates these two key factors. We evaluate Warble in a data set of tweets located in Barcelona during its festivities. The empirical results show that the model outperforms other state-of-the-art techniques in detecting various types of events while relying on a principled probabilistic framework that enables to reason under uncertainty. ; This work is partially supported by Obra Social "la Caixa", by the Spanish Ministry of Science and Innovation under contract (TIN2015-65316), by the Severo Ochoa Program (SEV2015-0493), by SGR programs of the Catalan Government (2014-SGR-1051, 2014-SGR-118), Collectiveware (TIN2015-66863-C2-1-R) and BSC/UPC NVIDIA GPU Center of Excellence.We would also like to thank the reviewers for their constructive feedback. ; Peer Reviewed ; Postprint (author's final draft)

Twitter has become one of the most popular Location-based Social Networks (LBSNs) that bridges physical and virtual worlds. Tweets, 140-character-long messages, are aimed to give answer to the What's happening? question. Occurrences and events in the real life (such as political protests, music concerts, natural disasters or terrorist acts) are usually reported through geo-located tweets by users on site. Uncovering event-related tweets from the rest is a challenging problem that necessarily requires exploiting different tweet features. With that in mind, we propose Tweet-SCAN, a novel event discovery technique based on the popular density-based clustering algorithm called DBSCAN. Tweet-SCAN takes into account four main features from a tweet, namely content, time, location and user to group together event-related tweets. The proposed technique models textual content through a probabilistic topic model called Hierarchical Dirichlet Process and introduces Jensen–Shannon distance for the task of neighborhood identification in the textual dimension. As a matter of fact, we show Tweet-SCAN performance in two real data sets of geo-located tweets posted during Barcelona local festivities in 2014 and 2015, for which some of the events were identified by domain experts beforehand. Through these tagged data sets, we are able to assess Tweet-SCAN capabilities to discover events, justify using a textual component and highlight the effects of several parameters. ; Peer Reviewed ; Postprint (author's final draft)

Twitter has become one of the most popular Location-based Social Networks (LBSNs) that bridges physical and virtual worlds. Tweets, 140-character-long messages, are aimed to give answer to the What's happening? question. Occurrences and events in the real life (such as political protests, music concerts, natural disasters or terrorist acts) are usually reported through geo-located tweets by users on site. Uncovering event-related tweets from the rest is a challenging problem that necessarily requires exploiting different tweet features. With that in mind, we propose Tweet-SCAN, a novel event discovery technique based on the popular density-based clustering algorithm called DBSCAN. Tweet-SCAN takes into account four main features from a tweet, namely content, time, location and user to group together event-related tweets. The proposed technique models textual content through a probabilistic topic model called Hierarchical Dirichlet Process and introduces Jensen–Shannon distance for the task of neighborhood identification in the textual dimension. As a matter of fact, we show Tweet-SCAN performance in two real data sets of geo-located tweets posted during Barcelona local festivities in 2014 and 2015, for which some of the events were identified by domain experts beforehand. Through these tagged data sets, we are able to assess Tweet-SCAN capabilities to discover events, justify using a textual component and highlight the effects of several parameters. ; Peer Reviewed ; Postprint (author's final draft)

The final publication is available at Springer via http://dx.doi.org/10.1007/s11042-020-09235-4 ; Image segmentation methods are usually trained with pixel-level annotations, which require significant human effort to collect. Weakly-supervised pipelines are the most common solution to address this constraint because they are trained with lower forms of supervision, such as bounding boxes or scribbles. Semi-supervised methods are another option, that leverage a large amount of unlabeled data and a limited number of strongly-labeled samples. In this second setup, samples to be strongly-annotated can be selected randomly or with an active learning mechanism that chooses the ones that will maximize the model performance. In this work, we propose a sample selection approach to decide which samples to annotate for semi-supervised instance segmentation. Our method consists in first predicting pseudo-masks for the unlabeled pool of samples, together with a score predicting the quality of each mask. This score is an estimate of the Intersection Over Union (IoU) of the segment with the ground truth mask. We study which samples should be annotated based on the quality score, leading to an improved performance for semi-supervised instance segmentation with low annotation budgets. ; This work was partially supported by the Spanish Ministry of Economy and Competitivity under contracts TIN2012-34557 by the BSC-CNS Severo Ochoa program (SEV-2011-00067), and contracts TEC2013-43935-R and TEC2016-75976-R. It has also been supported by grants 2014-SGR-1051 and 2014-SGR-1421 by the Government of Catalonia, and the European Regional Development Fund (ERDF). We would also like to acknowledge the valuable discussions with Victor Campos. ; Peer Reviewed ; Postprint (published version)

Recurrent Neural Networks (RNNs) continue to show outstanding performance in sequence modeling tasks. However, training RNNs on long sequences often face challenges like slow inference, vanishing gradients and difficulty in capturing long term dependencies. In backpropagation through time settings, these issues are tightly coupled with the large, sequential computational graph resulting from unfolding the RNN in time. We introduce the Skip RNN model which extends existing RNN models by learning to skip state updates and shortens the effective size of the computational graph. This model can also be encouraged to perform fewer state updates through a budget constraint. We evaluate the proposed model on various tasks and show how it can reduce the number of required RNN updates while preserving, and sometimes even improving, the performance of the baseline RNN models. Source code is publicly available at https://imatge-upc.github.io/skiprnn-2017-telecombcn/ ; This work was partially supported by the Spanish Ministry of Economy and Competitivity and the European Regional Development Fund (ERDF) under contracts TEC2016-75976-R and TIN2015- 65316-P, by the BSC-CNS Severo Ochoa program SEV-2015-0493, and grant 2014-SGR-1051 by the Catalan Government. Víctor Campos was supported by Obra Social "la Caixa" through La Caixa-Severo Ochoa International Doctoral Fellowship program. We would also like to thank the technical support team at the Barcelona Supercomputing Center. ; Postprint (published version)

Deep learning algorithms base their success on building high learning capacity models with millions of parameters that are tuned in a data-driven fashion. These models are trained by processing millions of examples, so that the development of more accurate algorithms is usually limited by the throughput of the computing devices on which they are trained. In this work, we explore how the training of a state-of-the-art neural network for computer vision can be parallelized on a distributed GPU cluster. The effect of distributing the training process is addressed from two different points of view. First, the scalability of the task and its performance in the distributed setting are analyzed. Second, the impact of distributed training methods on the final accuracy of the models is studied. ; This work is partially supported by the Spanish Ministry of Economy and Competitivity under contract TIN2012-34557, by the BSC-CNS Severo Ochoa program (SEV-2011-00067), by the SGR programmes (2014-SGR-1051 and 2014-SGR-1421) of the Catalan Government and by the framework of the project BigGraph TEC2013-43935-R, funded by the Spanish Ministerio de Economia y Competitividad and the European Regional Development Fund (ERDF). We also would like to thank the technical support team at the Barcelona Supercomputing center (BSC) especially to Carlos Tripiana. ; Peer Reviewed ; Postprint (published version)

Deep learning algorithms base their success on building high learning capacity models with millions of parameters that are tuned in a data-driven fashion. These models are trained by processing millions of examples, so that the development of more accurate algorithms is usually limited by the throughput of the computing devices on which they are trained. In this work, we explore how the training of a state-of-the-art neural network for computer vision can be parallelized on a distributed GPU cluster. The effect of distributing the training process is addressed from two different points of view. First, the scalability of the task and its performance in the distributed setting are analyzed. Second, the impact of distributed training methods on the final accuracy of the models is studied. ; This work is partially supported by the Spanish Ministry of Economy and Competitivity under contract TIN2012-34557, by the BSC-CNS Severo Ochoa program (SEV-2011-00067), by the SGR programmes (2014-SGR-1051 and 2014-SGR-1421) of the Catalan Government and by the framework of the project BigGraph TEC2013-43935-R, funded by the Spanish Ministerio de Economia y Competitividad and the European Regional Development Fund (ERDF). We also would like to thank the technical support team at the Barcelona Supercomputing center (BSC) especially to Carlos Tripiana. ; Peer Reviewed ; Postprint (published version)

One of the factors that have hindered progress in the areas of sign language recognition, translation, and production is the absence of large annotated datasets. Towards this end, we introduce How2Sign, a multimodal and multiview continuous American Sign Language (ASL) dataset, consisting of a parallel corpus of more than 80 hours of sign language videos and a set of corresponding modalities including speech, English transcripts, and depth. A three-hour subset was further recorded in the Panoptic studio enabling detailed 3D pose estimation. To evaluate the potential of How2Sign for real-world impact, we conduct a study with ASL signers and show that synthesized videos using our dataset can indeed be understood. The study further gives insights on challenges that computer vision should address in order to make progress in this field. Dataset website: http://how2sign.github.io/ ; This work received funding from Facebook through gifts to CMU and UPC; through projects TEC2016-75976-R, TIN2015- 65316-P, SEV-2015-0493 and PID2019-107255GB-C22 of the Spanish Government and 2017-SGR-1414 of Generalitat de Catalunya. This work used XSEDE's "Bridges" system at the Pittsburgh Supercomputing Center (NSF award ACI- 1445606). Amanda Duarte has received support from la Caixa Foundation (ID 100010434) under the fellowship code LCF/BQ/IN18/11660029. Shruti Palaskar was supported by the Facebook Fellowship program. ; Peer Reviewed ; Objectius de Desenvolupament Sostenible::10 - Reducció de les Desigualtats ; Objectius de Desenvolupament Sostenible::4 - Educació de Qualitat::4.5 - Per a 2030, eliminar les disparitats de gènere en l'educació i garantir l'accés en condicions d'igualtat a les persones vulnerables, incloses les persones amb discapacitat, els pobles indígenes i els nens i nenes en situacions de vulnerabilitat, a tots els nivells de l'ensenyament i la formació professional ; Objectius de Desenvolupament Sostenible::10 - Reducció de les Desigualtats::10.2 - Per a 2030, potenciar i promoure la inclusió social, econòmica i ...

Large scale time-dependent particle simulations can generate massive amounts of data, making it so that storing the results is often the slowest phase and the primary time bottleneck of the simulation. Furthermore, analysing this amount of data with traditional tools has become increasingly challenging, and it is often virtually impossible to have a visual representation of the full set. We propose a novel architecture that integrates an HPC-based multi-physics simulation code, a NoSQL database, and a data analysis and visualisation application. The goals are two: On the one hand, we aim to speed up the simulations taking advantage of the scalability of key-value data stores, while at the same time enabling real-time approximated data visualisation and interactive exploration. On the other hand, we want to make it efficient to explore and analyse the large data base of results produced. Therefore, this work represents a clear example of integrating High Performance Computing with High Performance Data Analytics. Our prototype proves the validity of our approach and shows great performance improvements. Indeed, we reduced by 67.5% the time to store the simulation while we made real-time queries run 52 times faster than alternative solutions. ; This work has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 720270 (HBP SGA1). It is also partially supported by the grant SEV-2011-00067 of Severo Ochoa Program, the TIN2015-65316-P project, with funding from the Spanish Ministry of Economy and Competitivity, the European Union FEDER funds, and the SGR 2014-SGR-1051. ; Peer Reviewed ; Postprint (published version)

Large scale time-dependent particle simulations can generate massive amounts of data, making it so that storing the results is often the slowest phase and the primary time bottleneck of the simulation. Furthermore, analysing this amount of data with traditional tools has become increasingly challenging, and it is often virtually impossible to have a visual representation of the full set. We propose a novel architecture that integrates an HPC-based multi-physics simulation code, a NoSQL database, and a data analysis and visualisation application. The goals are two: On the one hand, we aim to speed up the simulations taking advantage of the scalability of key-value data stores, while at the same time enabling real-time approximated data visualisation and interactive exploration. On the other hand, we want to make it efficient to explore and analyse the large data base of results produced. Therefore, this work represents a clear example of integrating High Performance Computing with High Performance Data Analytics. Our prototype proves the validity of our approach and shows great performance improvements. Indeed, we reduced by 67.5% the time to store the simulation while we made real-time queries run 52 times faster than alternative solutions. ; This work has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 720270 (HBP SGA1). It is also partially supported by the grant SEV-2011-00067 of Severo Ochoa Program, the TIN2015-65316-P project, with funding from the Spanish Ministry of Economy and Competitivity, the European Union FEDER funds, and the SGR 2014-SGR-1051. ; Peer Reviewed ; Postprint (published version)

The use of the Python programming language for scientific computing has been gaining momentum in the last years. The fact that it is compact and readable and its complete set of scientific libraries are two important characteristics that favour its adoption. Nevertheless, Python still lacks a solution for easily parallelizing generic scripts on distributed infrastructures, since the current alternatives mostly require the use of APIs for message passing or are restricted to embarrassingly parallel computations. In that sense, this paper presents PyCOMPSs, a framework that facilitates the development of parallel computational workflows in Python. In this approach, the user programs her script in a sequential fashion and decorates the functions to be run as asynchronous parallel tasks. A runtime system is in charge of exploiting the inherent concurrency of the script, detecting the data dependencies between tasks and spawning them to the available resources. Furthermore, we show how this programming model can be built on top of a Big Data storage architecture, where the data stored in the backend is abstracted and accessed from the application in the form of persistent objects. ; This work has been supported by the following institutions: the Spanish Government with grant SEV-2011-00067 of Severo Ochoa Program and contract Computaci´on de Altas Prestaciones VI (TIN2012-34557); by the SGR programme (2014-SGR-1051) of the Catalan Government; by the project The Human Brain Project, funded by the European Commission under contract 604102; and by the Intel-BSC Exascale Lab collaboration. ; Peer Reviewed ; Postprint (author's final draft)