This paper describes the work done by Priberam in the development of a question answering (QA) system for Portuguese. The system was built using the company’s natural language processing (NLP) workbench and information retrieval technology. Special focus is given to question analysis, document and sentence retrieval, as well as answer extraction stages. The paper discusses the system’s performance in the context of the QA@CLEF 2005 evaluation.

/pdf/priberam-s-question-answering-system-for-portuguese-338e3iu1qp.pdf

Priberam’s question answering system for portuguese

A growing number of commercial satellite companies provide easily accessible satellite imagery. Overhead imagery is used by numerous industries including agriculture, forestry, natural disaster analysis, and meteorology. Satellite images, just as any other images, can be tampered with image manipulation tools. Manipulation detection methods created for images captured by "consumer cameras" tend to fail when used on satellite images due to the differences in image sensors, image acquisition, and processing. In this paper we propose an unsupervised technique that uses a Vision Transformer to detect spliced areas within satellite images. We introduce a new dataset which includes manipulated satellite images that contain spliced objects. We show that our proposed approach performs better than existing unsupervised splicing detection techniques.

/pdf/manipulation-detection-in-satellite-images-using-vision-5g8ub3fb72.pdf

Manipulation Detection in Satellite Images Using Vision Transformer

Modeling and control in open-channel irrigation systems: A review

The systematic monitoring of the Earth using optical satellites is limited by the presence of clouds. Accurately detecting these clouds is necessary to exploit satellite image archives in remote sensing applications. Despite many developments, cloud detection remains an unsolved problem with room for improvement, especially over bright surfaces and thin clouds. Recently, advances in cloud masking using deep learning have shown significant boosts in cloud detection accuracy. However, these works are validated in heterogeneous manners, and the comparison with operational threshold-based schemes is not consistent among many of them. In this work, we systematically compare deep learning models trained on Landsat-8 images on different Landsat-8 and Sentinel-2 publicly available datasets. Overall, we show that deep learning models exhibit a high detection accuracy when trained and tested on independent images from the same Landsat-8 dataset (intra-dataset validation), outperforming operational algorithms. However, the performance of deep learning models is similar to operational threshold-based ones when they are tested on different datasets of Landsat-8 images (inter-dataset validation) or datasets from a different sensor with similar radiometric characteristics such as Sentinel-2 (cross-sensor validation). The results suggest that (i) the development of cloud detection methods for new satellites can be based on deep learning models trained on data from similar sensors and (ii) there is a strong dependence of deep learning models on the dataset used for training and testing, which highlights the necessity of standardized datasets and procedures for benchmarking cloud detection models in the future.

Benchmarking Deep Learning Models for Cloud Detection in Landsat-8 and Sentinel-2 Images

This paper presents a portable device for outdoor air quality measurement that provides concentration values for the main pollutants: NO2, NO, CO, O3, PM2.5 and PM10, and other values such as temperature, humidity, location, and date. The device is based on the use of commercial electrochemical gas and optical particle matter sensors with a careful design of the electronics for reducing the electrical noise and increasing the accuracy of the measurements. The result is a low-cost system with IoT technology that connects to the Internet through a GSM module and sends all real-time data to a cloud platform with storage and computational potential. Two identical devices were fabricated and installed on a mobile reference measurement unit and deployed in Badajoz, Spain. The results of a two-month field campaign are presented and published. Data obtained from these measurements were calibrated using linear regression and neural network techniques. Good performance has been achieved for both gaseous pollutants (with a Pearson correlation coefficient of up to 0.97) and PM sensors.

Low-Cost Air Quality Measurement System Based on Electrochemical and PM Sensors with Cloud Connection.

Given the continuous increase in the global population, the food manufacturers are advocated to either intensify the use of cropland or expand the farmland, making land cover and land usage dynamics mapping vital in the area of remote sensing. In this regard, identifying and classifying a high-resolution satellite imagery scene is a prime challenge. Several approaches have been proposed either by using static rule-based thresholds (with limitation of diversity) or neural network (with data-dependent limitations). This paper adopts the inductive approach to learning from surface reflectances. A manually labeled Sentinel-2 dataset was used to build a Machine Learning (ML) model for scene classification, distinguishing six classes (Water, Shadow, Cirrus, Cloud, Snow, and Other). This models was accessed and further compared to the European Space Agency (ESA) Sen2Cor package. The proposed ML model presents a Micro-F1 value of 0.84, a considerable improvement when compared to the Sen2Cor corresponding performance of 0.59. Focusing on the problem of optical satellite image scene classification, the main research contributions of this paper are: (a) an extended manually labeled Sentinel-2 database adding surface reflectance values to an existing dataset; (b) an ensemble-based and a Neural-Network-based ML models; (c) an evaluation of model sensitivity, biasness, and diverse ability in classifying multiple classes over different geographic Sentinel-2 imagery, and finally, (d) the benchmarking of the ML approach against the Sen2Cor package.

Sentinel-2 Image Scene Classification: A Comparison between Sen2Cor and a Machine Learning Approach

The approach followed by the University of Evora team in order to build a system able to participate in the QA-CLEF task is described. The system is based in two steps: for each question, a first information retrieval task selects a set of potentially relevant documents; then, each of these documents is analysed trying to obtain their semantic representation and the answer to the initial query. The proposed approach was applied to the test set of the QA-CLEF for the Portuguese language and the obtained results were quite interesting and motivating and allowed the identification of strong and weak characteristics of the system.

/pdf/the-university-of-evora-approach-to-qa-clef-2004-3qv3k2jgqh.pdf

The University of Évora approach to QA@CLEF-2004.

In this work we present NeMODe a declarative system for Computer Network Intrusion detection which provides a declarative Domain Specific Language for describing computer network intrusion signatures that could spread across several network packets, which allows to state constraints over network packets, describing relations between several packets, and providing several back-end detection mechanisms which relies on Constraint Programming (CP) methodologies to find those intrusions.

/pdf/using-constraints-for-intrusion-detection-the-nemode-system-3k9s67v7ff.pdf

Using constraints for intrusion detection: the NeMODe system

Current technology facilitates access to the vast amount of information that is produced every day. Both individuals and companies are active consumers of data from the Web and other sources, and these data guide decision making. Due to the huge volume of data to be processed in a business context, managers rely on decision support systems to facilitate data analysis. OLAP tools are Business Intelligence solutions for multidimensional analysis of data, allowing the user to control the perspective and the degree of detail in each dimension of the analysis. A conventional OLAP system is configured to a set of analysis scenarios associated with multidimensional data cubes in the repository. To handle a more spontaneous query, not supported in these provided scenarios, one must have specialized technical skills in data analytics. This makes it very difficult for average users to be autonomous in analyzing their data, as they will always need the assistance of specialists. This article describes an ontology-based natural language interface whose goal is to simplify and make more flexible and intuitive the interaction between users and OLAP solutions. Instead of programming an MDX query, the user can freely write a question in his own human language. The system interprets this question by combining the requested information elements, and generates an answer from the OLAP repository.

/pdf/binli-an-ontology-based-natural-language-interface-for-2o0jmqjlto.pdf

BINLI: An Ontology-Based Natural Language Interface for Multidimensional Data Analysis

The interest in the quality of air is likely to increase, as the public concern for health and environmental issues is on the rise. So far, most data available comes from a small numbers of government-owned sensors, lacking a wide coverage of the entire reality. Improving the amount of data available thus involves reducing the cost of sensors and make their readings accessible to the public. The NanoSen-AQM project aims to do precisely that. Create and use vast numbers of low-cost nano-sensors, to make their data accessible for the public. To achieve such an ambitious goal, the project will use state-of-the-art techniques from Machine Learning and mobile and web development frameworks. As a result, the NanoSen-AQM platform should provide free access to the public and low-cost of entry for sensor owners willing to share their data.

Pedro Salgueiro

Papers

Sentinel-2 Image Scene Classification: A Comparison between Sen2Cor and a Machine Learning Approach

The University of Évora approach to QA@CLEF-2004.

Using constraints for intrusion detection: the NeMODe system

BINLI: An Ontology-Based Natural Language Interface for Multidimensional Data Analysis

An Online Platform For Real-Time Air Quality Monitoring