Search

Time-series forecasting is one of the most active research topics in artificial intelligence. Applications in real-world time series should consider two factors for achieving reliable predictions: modeling dynamic dependencies among multiple variables and adjusting the model's intrinsic hyperparameters. A still open gap in that literature is that statistical and ensemble learning approaches systematically present lower predictive performance than deep learning methods. They generally disregard the data sequence aspect entangled with multivariate data represented in more than one time series. Conversely, this work presents a novel neural network architecture for time-series forecasting that combines the power of graph evolution with deep recurrent learning on distinct data distributions; we named our method Recurrent Graph Evolution Neural Network (ReGENN). The idea is to infer multiple multivariate relationships between co-occurring time-series by assuming that the temporal data depends not only on inner variables and intra-temporal relationships (i.e., observations from itself) but also on outer variables and inter-temporal relationships (i.e., observations from other-selves). An extensive set of experiments was conducted comparing ReGENN with dozens of ensemble methods and classical statistical ones, showing sound improvement of up to 64.87% over the competing algorithms. Furthermore, we present an analysis of the intermediate weights arising from ReGENN, showing that by looking at inter and intra-temporal relationships simultaneously, time-series forecasting is majorly improved if paying attention to how multiple multivariate data synchronously evolve., Comment: IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021

International audience; The interest for patient trajectory prediction, a sort of computer-aided medicine, has steadily increased with the pace of artificial intelligence innovation. Notwithstanding, the design of effective systems able to predict clinical outcomes based on the history of a patient is far from trivial. Works so far are based on neural architectures with low performance, especially when using low-cardinality datasets; alternatively, complex inference approaches are hard to reproduce and/or extrapolate as they are designed for very specific circumstances. We introduce LIG-Doctor, an artificial neural network architecture based on two Minimal Gated Recurrent Unit networks functioning in a bidirectional parallel manner, benefiting from temporal events both forward and backward. In comparison to state-of-the-art works, consistent improvements were achieved in prognosis prediction, as assessed with metrics Recall@k, Precision@k, F1-score, and AUC-ROC. Besides the detailed delineation of our architecture, a sequence of experiments is reported with insights that progressively guided design decisions to inspire future works on similar problems. Our results shall contribute to the improvement of computer-aided medicine and, more generally, to processes related to the design of neural network architectures.

Human mobility has a significant impact on several layers of society, from infrastructural planning and economics to the spread of diseases and crime. Representing the system as a complex network, in which nodes are assigned to regions (e.g., a city) and links indicate the flow of people between two of them, physics-inspired models have been proposed to quantify the number of people migrating from one city to the other. Despite the advances made by these models, our ability to predict the number of commuters and reconstruct mobility networks remains limited. Here, we propose an alternative approach using machine learning and 22 urban indicators to predict the flow of people and reconstruct the intercity commuters network. Our results reveal that predictions based on machine learning algorithms and urban indicators can reconstruct the commuters network with 90.4% of accuracy and describe 77.6% of the variance observed in the flow of people between cities. We also identify essential features to recover the network structure and the urban indicators mostly related to commuting patterns. As previously reported, distance plays a significant role in commuting, but other indicators, such as Gross Domestic Product (GDP) and unemployment rate, are also driven-forces for people to commute. We believe that our results shed new lights on the modeling of migration and reinforce the role of urban indicators on commuting patterns. Also, because link-prediction and network reconstruction are still open challenges in network science, our results have implications in other areas, like economics, social sciences, and biology, where node attributes can give us information about the existence of links connecting entities in the network., Comment: 28 pages, 5 figures

Processing navigational graph-like queries in relational databases requires executing several recursive join operations, which are computationally costly. However, when the need for graph-like queries arises, applications often execute a sequence of related queries in a single session. We argue that it is possible to reduce the total cost of a set of related queries, by expanding individual intermediate results and sharing them among multiple queries. SHARq is our framework that enables sharing intermediate results of the common graph-like queries Single-Source Shortest Paths (SSSP), Connected Components (CC), and PageRank (PR). Our solution prepares result tables expanded with additional columns to store partial results of graph-like query combinations, such as multiple SSSP, or a sequence of queries comprising SSSP, CC, and PR. Experimental results on 9 datasets show query speedups of up to ten times when combining multiple SSSP queries, and up to two times when combining SSSP, CC, and PR queries. The results reveal a significant reduction in the query time, providing timely results for analyses relying on multiple navigational graph-like queries.

Made available in DSpace on 2022-04-28T19:40:22Z (GMT). No. of bitstreams: 0 Previous issue date: 2021-07-15 Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) In exploratory tasks involving high-dimensional datasets, dimensionality reduction (DR) techniques help analysts to discover patterns and other useful information. Although scatter plot representations of DR results allow for cluster identification and similarity analysis, such a visual metaphor presents problems when the number of instances of the dataset increases, resulting in cluttered visualizations. In this work, we propose a scatter plot-based multilevel approach to display DR results and address clutter-related problems when visualizing large datasets, together with the definition of a methodology to use focus+context interaction on non-hierarchical embeddings. The proposed technique, called ExplorerTree, uses a sampling selection technique on scatter plots to reduce visual clutter and guide users through exploratory tasks. We demonstrate ExplorerTree's effectiveness through a use case, where we visually explore activation images of the convolutional layers of a neural network. Finally, we also conducted a user experiment to evaluate ExplorerTree's ability to convey embedding structures using different sampling strategies. Faculty of Sciences and Technology São Paulo State University (UNESP) Faculty of Computer Science Dalhousie University Institute of Mathematics and Computer Sciences University of São Paulo Faculty of Sciences and Technology São Paulo State University (UNESP) FAPESP: 2016/11707-6 FAPESP: 2017/17450-0 FAPESP: 2018/17881-3 FAPESP: 2018/25755-8

Early identification of the type of skin lesion, some of them carcinogenic, is of paramount importance. Currently, the use of Convolutional Neural Networks (CNNs) is the mainline of investigation for the automated analysis of such lesions. Most of the existing works, however, were designed by transfer learning general-purpose CNN architectures, adapting existing methods to the domain of dermatology. Despite effective, this approach poses inflexibility and high processing costs. In this work, we introduce a novel architecture that benefits from cutting-edge CNN techniques Aggregated Transformations combined to the mechanism of Squeeze-and-Excite organized in a residual block; our architecture is designed and trained from scratch to solve both the binary melanoma detection problem, as well as the multi-class skin-lesion classification problem. Our results demonstrate that such an architecture is competitive to major state-of-the-art architectures adapted to the domain of skin-lesion diagnosis. Our architecture is prone to evolve and to provide low processing cost for real-world in situ applications using a much smaller number of weights if compared to previous works.

As more data is generated from medical attendances and as Artificial Neural Networks gain momentum in research and industry, computer-aided medical prognosis has become a promising technology. A common approach to perform automated prognoses relies on textual clinical notes extracted from Electronic Health Records (EHRs). Data from EHRs are fed to neural networks that produce a set with the most probable medical problems to which a patient is subject in her/his clinical future, including clinical conditions, mortality, and readmission. Following this research line, we introduce a methodology that takes advantage of the unstructured text found in clinical notes by applying preprocessing, concepts extraction, and fine-tuned neural networks to predict the most probable medical problems to follow in a patient’s clinical trajectory. Different from former works that focus on word embeddings and raw sets of extracted concepts, we generate a refined set of Unified Medical Language System (UMLS) concepts by applying a similarity threshold filter and a list of acceptable concept types. In our prediction experiments, our method demonstrated AUC-ROC performance of 0.91 for diagnosis codes, 0.93 for mortality, and 0.72 for readmission, determining an efficacy that rivals state-of-the-art works. Our findings contribute to the development of automated prognosis systems in hospitals where text is the main source of clinical history.

Pesticides have been heavily used in the cultivation of major crops, contributing to the increase of crop production over the past decades. However, in many cases their appropriate use and calibration of machines rely upon dated evaluation methodologies that cannot precisely estimate how well the pesticides’ are being applied to the crop. A few strategies have been proposed in former works, yet their elevated costs and low portability do not permit their wide spread adoption. This work introduces and experimentally assesses a novel tool that functions as a smartphone-based mobile application, named DropLeaf - Spraying Meter. Tests performed using DropLeaf demonstrated that, notwithstanding its simplicity, it can estimate the pesticide coverage with high precision. Our methodology is based on the development of custom image analysis software for real-time assessment of spraying deposition of water-sensitive papers. The proposed tool can be extensively used by farmers and agronomists carrying regular smartphones, improving the utilization of pesticides with well-being, ecological, and monetary advantages. DropLeaf can be easily used for spray drift assessment of different methods, including emerging unmanned aerial vehicle and smart sprayers.

Made available in DSpace on 2019-10-04T12:35:15Z (GMT). No. of bitstreams: 0 Previous issue date: 2019-01-01 Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) We propose an algebraic tool-set and related algorithms to track access problems not obviously observed in urban environments represented as street mashes. Our tool-set assumes that points of interest must be promptly accessible within the paths of a street network. Over that assumption, we introduce formalisms and computational tools to detect and evaluate access-related problems. The output of our methods, in the form of inconsistent nodes found in a given city, has the potential to assist the decision-making process regarding the positioning of resources, building of new services, and outlining the initial design of a city. (C) 2018 Elsevier B.V. All rights reserved. Univ Sao Paulo, Sao Carlos, SP, Brazil Univ Fed Mato Grosso do Sul, Comp Sci Dept, Ponta Pora, MS, Brazil Sao Paulo State Univ, Dept Math & Computat, Presidente Prudente, SP, Brazil Sao Paulo State Univ, Dept Math & Computat, Presidente Prudente, SP, Brazil CNPq: 167967/2017-7 FAPESP: 2016/17078-0 FAPESP: 2016/02557-0 FAPESP: 2017/08376-0

Recursive queries are one of the main mechanisms in Relational Database Management Systems to process topology-aware, or graph-like, queries. However, existing works focus only on optimizing the recursive query statements and processing, disregarding the potential physical arrangements that might improve performance. In this work, we propose to use an approach based on adjacent-list storage to physically organize the graph-like data aiming at both reducing the recursive query time and the number of I/O operations. By using Clustered Tables, we tied the adjacency list in chunks for (i) storing both vertex and edge tables together in a Combined Tables approach; and (ii) reordering the edge table with the Edge Clustered Table approach using 20% and 80% of the total adjacency list size. The clustered approaches enabled a faster recursive query processing (up to 22%) and a reduction of up to 61% in the number of page accesses when compared to the Conventional approach. When starting from multiple vertices, the Combined Tables approach achieved a query reduction time of up to 50% in the first join operation, and Edge Clustered Table 20% provided an overall time reduction of up to 20%. The results show that our physical design is effective and allows one to use recursive queries without adaptations.

What if a successful company starts to receive a torrent of low-valued (one or two stars) recommendations in its mobile apps from multiple users within a short (say one month) period of time? Is it legitimate evidence that the apps have lost in quality, or an intentional plan (via lockstep behavior) to steal market share through defamation? In the case of a systematic attack to one's reputation, it might not be possible to manually discern between legitimate and fraudulent interaction within the huge universe of possibilities of user-product recommendation. Previous works have focused on this issue, but none of them took into account the context, modeling, and scale that we consider in this paper. Here, we propose the novel method Online-Recommendation Fraud ExcLuder (ORFEL) to detect defamation and/or illegitimate promotion of online products by using vertex-centric asynchronous parallel processing of bipartite (users-products) graphs. With an innovative algorithm, our results demonstrate both efficacy and efficiency -- over 95% of potential attacks were detected, and ORFEL was at least two orders of magnitude faster than the state-of-the-art. Over a novel methodology, our main contributions are: (1) a new algorithmic solution; (2) one scalable approach; and (3) a novel context and modeling of the problem, which now addresses both defamation and illegitimate promotion. Our work deals with relevant issues of the Web 2.0, potentially augmenting the credibility of online recommendation to prevent losses to both customers and vendors., Comment: http://www.sciencedirect.com/science/article/pii/S0020025516307320. Information Sciences, Available online 5 September 2016, ISSN 0020-0255

This paper presents the extract-transform-and-load (ETL) process from the Electronic Patient Records (ePR) at the Heart Institute (InCor) to the OMOP Common Data Model (CDM) format. We describe the initial database characterization, relational source mappings, selection filters, data transformations and patient de-identification using the open-source OHDSI tools and SQL scripts. We evaluate the resulting InCor-CDM database by recreating the same patient cohort from a previous reference study (over the original data source) and comparing the cohorts' descriptive statistics and inclusion reports. The results exhibit that up to 91% of the reference patients were retrieved by our method from the ePR through InCor-CDM, with AUC=0.938. The results indicate that the method that we employed was able to produce a new database that was both consistent with the original data and in accordance to the OMOP CDM standard.

Street networks are complex networks that represent the topology and geometry of cities; as so, they can be used to solve problems related to ill-designed urban structures. This application, in real urban scenarios, has been the focus of several types of research, from cities characterization to transportation enhancement. Nevertheless, these works lack a clear and in-depth methodology to characterize the urban space by means of complex networks. Aided by topo-geometrical measures from street networks, we present a methodology to identify what we call urban inconsistencies, which are characterized by low-access regions containing nodes (crossing streets) that lack efficient access from or to other regions in a city. We devised algorithms capable of preprocessing and analyzing street networks, pointing to existing mobility problems in a city. We identify inconsistencies that pertain to a given node where a facility of interest is currently placed; the results introduce ways to assist in the urban planning and design processes. Our techniques are discussed through the analysis of a real-world city. In a real context, the methods provide basis for analyzing and improving the placement of facilities.

Complex networks are commonly used to model urban street networks, which allows aiding the analysis of criminal activities in cities. Despite several works focusing on such application, there is a lack of a clear methodology focused in the analysis of crime behavior. In this sense, we propose a methodology for employing complex networks in the analysis of criminality spread within criminal areas of a city. Here, we evaluate synthetic cases of crime propagation concerning real criminal data from the North American city of San Francisco — CA. Our results confirm the effectiveness of our methodology in analyzing the crime behavior by means of criminality spread. Hence, this paper renders further development and planning on public safety in cities.

Soybean is one of the ten greatest crops in the world, answering for billion-dollar businesses every year. This crop suffers from insect herbivory that costs millions from producers. Hence, constant monitoring of the crop foliar damage is necessary to guide the application of insecticides. However, current methods to measure foliar damage are expensive and dependent on laboratory facilities, in some cases, depending on complex devices. To cope with these shortcomings, we introduce an image processing methodology to measure the foliar damage in soybean leaves. We developed a non-destructive imaging method based on two techniques, Otsu segmentation and Bezier curves, to estimate the foliar loss in leaves with or without border damage. We instantiate our methodology in a mobile application named BioLeaf, which is freely distributed for smartphone users. We experimented with real-world leaves collected from a soybean crop in Brazil. Our results demonstrated that BioLeaf achieves foliar damage quantification with precision comparable to that of human specialists. With these results, our proposal might assist soybean producers, reducing the time to measure foliar damage, reducing analytical costs, and defining a commodity application that is applicable not only to soy, but also to different crops such as cotton, bean, potato, coffee, and vegetables.

As redes complexas contribuem para a pesquisa computacional por sua capacidade de projetar sistemas modelados por vértices e arestas. Eles fornecem meios para descrever estruturas urbanas por meio das malhas viárias, expressando predicados que se referem ao fluxo e ao transporte em zonas urbanas. Este trabalho tem o objetivo de descrever as interações entre diferentes cidades usando seus vetores de características pela análise de informações viárias e das métricas inerentes aos seus elementos. Propõe-se uma análise baseada no uso de mapas digitais, pois permitem abordagens para modelagem de dados e extração de características que suportam atividades analíticas. Os resultados deste trabalho são baseados na análise de 645 cidades, que formam o estado brasileiro de São Paulo; tais resultados demonstram como características extraídas por métricas de grafos descrevem indicadores urbanos que estão enraizados na topologia da rede, e como podem revelar diferenças entre cidades distintas.

The need for higher agricultural productivity has demanded the intensive use of pesticides. However, their correct use depends on assessment methods that can accurately predict how well the pesticides' spraying covered the intended crop region. Some methods have been proposed in the literature, but their high cost and low portability harm their widespread use. This paper proposes and experimentally evaluates a new methodology based on the use of a smartphone-based mobile application, named DropLeaf. Experiments performed using DropLeaf showed that, in addition to its versatility, it can predict with high accuracy the pesticide spraying. DropLeaf is a five-fold image-processing methodology based on: (i) color space conversion, (ii) threshold noise removal, (iii) convolutional operations of dilation and erosion, (iv) detection of contour markers in the water-sensitive card, and, (v) identification of droplets via the marker-controlled watershed transformation. The authors performed successful experiments over two case studies, the first using a set of synthetic cards and the second using a real-world crop. The proposed tool can be broadly used by farmers equipped with conventional mobile phones, improving the use of pesticides with health, environmental and financial benefits., Comment: Paper to be published on the 33rd ACM/SIGAPP Symposium On Applied Computing (SAC), 2018

Complex networks have been widely used in science and technology because of their ability to represent several systems. One of these systems is found in Biochemistry, in which the synthesis of new nanoparticles is a hot topic. However, the interpretation of experimental results in the search of new nanoparticles poses several challenges. This is due to the characteristics of nanoparticles images and due to their multiple intricate properties; one property of recurrent interest is the agglomeration of particles. Addressing this issue, this paper introduces an approach that uses complex networks to detect and describe nanoparticle agglomerates so to foster easier and more insightful analyses. In this approach, each detected particle in an image corresponds to a vertice and the distances between the particles define a criterion for creating edges. Edges are created if the distance is smaller than a radius of interest. Once this network is set, we calculate several discrete measures able to reveal the most outstanding agglomerates in a nanoparticle image. Experimental results using images of Scanning Tunneling Microscopy (STM) of gold nanoparticles demonstrated the effectiveness of the proposed approach over several samples, as reflected by the separability between particles in three usual settings. The results also demonstrated efficacy for both convex and non-convex agglomerates., Comment: 10 pages

Relational Database Management Systems (RDBMSs) are widely employed in several applications, including those that deal with data modeled as graphs. Existing solutions store every edge in a distinct row in the edge table, however, for most cases, such modeling does not provide adequate performance. In this work, we propose Edge-k, a technique to group the vertex neighborhood into a reduced number of rows in a table through additional columns that stores up to k edges per row. The technique provides a better table organization and reduces both table size and query processing time. We evaluate Edge-k table management for insert, update, delete and bulkload operations, and compare the query processing performance both with the conventional edge table — adopted by the existing frameworks — and with the Neo4j graph database. Experiments using Single-Source Shortest Path (SSSP) queries reveal that our new proposal approach always outperforms the conventional edge table as well as it was faster than Neo4j for the first iterations, being slightly slower than Neo4j only for iterations after having loaded the whole graph from disk to memory. It was able to reach a speedup of 66% over a representative real dataset, with an average reduction of up to 58% in our tests. The average speedup over synthetic datasets was up to 54%. Edge-k was also the best one when performing graph degree distribution queries. Moreover, the Edge-k table obtained a processing time reduction of 70% for bulkload operations, despite having an overhead of 50% for individual insert, update and delete operations. Finally, Edge-k advances the state of the art for graph data management within relational database systems.

Crowdsourcing solutions can be helpful to extract information from disaster-related data during crisis management. However, certain information can only be obtained through similarity operations. Some of them also depend on additional data stored in a Relational Database Management System (RDBMS). In this context, several works focus on crisis management supported by data. Nevertheless, none of them provide a methodology for employing a similarity-enabled RDBMS in disaster-relief tasks. To fill this gap, we introduce a methodology together with the Data-Centric Crisis Management (DCCM) architecture, which employs our methods over a similarity-enabled RDBMS. We evaluate our proposal through three tasks: classification of incoming data regarding current events, identifying relevant information to guide rescue teams; filtering of incoming data, enhancing the decision support by removing near-duplicate data; and similarity retrieval of historical data, supporting analytical comprehension of the crisis context. To make it possible, similarity-based operations were implemented within one popular, open-source RDBMS. Results using real data from Flickr show that our proposal is feasible for real-time applications. In addition to high performance, accurate results were obtained with a proper combination of techniques for each task. Hence, we expect our work to provide a framework for further developments on crisis management solutions.