Recommending media content based on machine learning methods

Dissertação para obtenção do Grau de Mestre em Engenharia Informática

ZoomAzores project: implementation of a WebGIS for Nature and Adventure Tourism.

Dissertação apresentada como requisito parcial para obtenção do grau de Mestre em Estatística e Gestão de Informação

The clothing industry as a globalized sector: implications for work organisation, quality of work and job content

The clothing sector in several countries is still seen, in many aspects as a traditional sector with some average characteristics, nevertheless is a very important sector in terms of labour market. Globalization and de-localization are having a strong impact in the organisation of work and in occupational careers. Very few companies are able to keep a position in the market without changes in organisation of work and workers, founding different ways to face this reality according to size, capital and position. We could find two main paths: one where companies outsource production to another territory, close and/ or dismissal the workers; other path, where companies up skilled their capacities. This paper will present some results from the European project WORKS – Work organisation and restructuring in the knowledge society (6th Framework Programme), focusing the Portuguese case studies in several clothing companies in a comparative analysis with some other European countrie

Combining open and closed world reasoning for the semantic web

Dissertação para obtenção do Grau de Doutor em Informática

Brokerage Platform for Media Content Recommendation

Near real time media content personalisation is nowadays a major challenge involving media content sources, distributors and viewers. This paper describes an approach to seamless recommendation, negotiation and transaction of personalised media content. It adopts an integrated view of the problem by proposing, on the business-to-business (B2B) side, a brokerage platform to negotiate the media items on behalf of the media content distributors and sources, providing viewers, on the business-to-consumer (B2C) side, with a personalised electronic programme guide (EPG) containing the set of recommended items after negotiation. In this setup, when a viewer connects, the distributor looks up and invites sources to negotiate the contents of the viewer personal EPG. The proposed multi-agent brokerage platform is structured in four layers, modelling the registration, service agreement, partner lookup, invitation as well as item recommendation, negotiation and transaction stages of the B2B processes. The recommendation service is a rule-based switch hybrid filter, including six collaborative and two content-based filters. The rule-based system selects, at runtime, the filter(s) to apply as well as the final set of recommendations to present. The filter selection is based on the data available, ranging from the history of items watched to the ratings and/or tags assigned to the items by the viewer. Additionally, this module implements (i) a novel item stereotype to represent newly arrived items, (ii) a standard user stereotype for new users, (iii) a novel passive user tag cloud stereotype for socially passive users, and (iv) a new content-based filter named the collinearity and proximity similarity (CPS). At the end of the paper, we present off-line results and a case study describing how the recommendation service works. The proposed system provides, to our knowledge, an excellent holistic solution to the problem of recommending multimedia contents.

Platform for Smart Car to Car Content Delivery: Results of CISTER Research Centre within CarCoDe project

Poster presented in Redes de Veiculos nas sociedades do futuro (RVSF 2015). 3, Jun, 2015. Castelo Branco, Portugal.

Design a computer programming learning environment for Massive Open Online Courses

Teaching and learning computer programming is as challenging as difficult. Assessing the work of students and providing individualised feedback to all is time-consuming and error prone for teachers and frequently involves a time delay. The existent tools and specifications prove to be insufficient in complex evaluation domains where there is a greater need to practice. At the same time Massive Open Online Courses (MOOC) are appearing revealing a new way of learning, more dynamic and more accessible. However this new paradigm raises serious questions regarding the monitoring of student progress and its timely feedback. This paper provides a conceptual design model for a computer programming learning environment. This environment uses the portal interface design model gathering information from a network of services such as repositories and program evaluators. The design model includes also the integration with learning management systems, a central piece in the MOOC realm, endowing the model with characteristics such as scalability, collaboration and interoperability. This model is not limited to the domain of computer programming and can be adapted to any complex area that requires systematic evaluation with immediate feedback.

Effects of Soil Compaction and Organic Carbon Content on Preferential Flow in Loamy Field Soils

Abstract: Preferential flow and transport through macropores affect plant water use efficiency and enhance leaching of agrochemicals and the transport of colloids, thereby increasing the risk for contamination of groundwater resources. The effects of soil compaction, expressed in terms of bulk density (BD), and organic carbon (OC) content on preferential flow and transport were investigated using 150 undisturbed soil cores sampled from 15 × 15–m grids on two field sites. Both fields had loamy textures, but one site had significantly higher OC content. Leaching experiments were conducted in each core by applying a constant irrigation rate of 10 mm h−1 with a pulse application of tritium tracer. Five percent tritium mass arrival times and apparent dispersivities were derived from each of the tracer breakthrough curves and correlated with texture, OC content, and BD to assess the spatial distribution of preferential flow and transport across the investigated fields. Soils from both fields showed strong positive correlations between BD and preferential flow. Interestingly, the relationships between BD and tracer transport characteristics were markedly different for the two fields, although the relationship between BD and macroporosity was nearly identical. The difference was likely caused by the higher contents of fines and OC at one of the fields leading to stronger aggregation, smaller matrix permeability, and a more pronounced pipe-like pore system with well-aligned macropores.

Tuning metadata for better movie content-based recommendation systems

The increasing number of television channels, on-demand services and online content, is expected to contribute to a better quality of experience for a costumer of such a service. However, the lack of efficient methods for finding the right content, adapted to personal interests, may lead to a progressive loss of clients. In such a scenario, recommendation systems are seen as a tool that can fill this gap and contribute to the loyalty of users. Multimedia content, namely films and television programmes are usually described using a set of metadata elements that include the title, a genre, the date of production, and the list of directors and actors. This paper provides a deep study on how the use of different metadata elements can contribute to increase the quality of the recommendations suggested. The analysis is conducted using Netflix and Movielens datasets and aspects such as the granularity of the descriptions, the accuracy metric used and the sparsity of the data are taken into account. Comparisons with collaborative approaches are also presented.

A Multi–Agent Brokerage Platform for Media Content Recommendation

Near real time media content personalisation is nowadays a major challenge involving media content sources, distributors and viewers. This paper describes an approach to seamless recommendation, negotiation and transaction of personalised media content. It adopts an integrated view of the problem by proposing, on the business-to-business (B2B) side, a brokerage platform to negotiate the media items on behalf of the media content distributors and sources, providing viewers, on the business-to-consumer (B2C) side, with a personalised electronic programme guide (EPG) containing the set of recommended items after negotiation. In this setup, when a viewer connects, the distributor looks up and invites sources to negotiate the contents of the viewer personal EPG. The proposed multi-agent brokerage platform is structured in four layers, modelling the registration, service agreement, partner lookup, invitation as well as item recommendation, negotiation and transaction stages of the B2B processes. The recommendation service is a rule-based switch hybrid filter, including six collaborative and two content-based filters. The rule-based system selects, at runtime, the filter(s) to apply as well as the final set of recommendations to present. The filter selection is based on the data available, ranging from the history of items watched to the ratings and/or tags assigned to the items by the viewer. Additionally, this module implements (i) a novel item stereotype to represent newly arrived items, (ii) a standard user stereotype for new users, (iii) a novel passive user tag cloud stereotype for socially passive users, and (iv) a new content-based filter named the collinearity and proximity similarity (CPS). At the end of the paper, we present off-line results and a case study describing how the recommendation service works. The proposed system provides, to our knowledge, an excellent holistic solution to the problem of recommending multimedia contents.

Chronic hepatitis C: hepatic iron content does not correlate with response to antiviral therapy

The complex interaction between hepatitis C virus infection, iron homeostasis and the response to antiviral treatment remains controversial. The aim of this study was to evaluate the influence of hepatic iron concentration (HIC) on the sustained virological response (SVR) to antiviral therapy in patients with chronic hepatitis C. A total of 50 patients who underwent pretreatment liver biopsy with assessment of HIC by graphite furnace atomic absorption spectroscopy and were subsequently submitted to antiviral treatment with interferon/peginterferon and ribavirin were included in the study. Patients with alcoholism, history of multiple blood transfusion, chronic kidney disease, hemolytic anemia and parenteral iron therapy were excluded. The iron related markers and HIC were compared between those who achieved an SVR and non-responders (NR) patients. The mean age was 45.7 years and the proportion of patients' gender was not different between SVR and NR patients. The median serum iron was 138 and 134 µg/dL (p = 0.9), the median serum ferritin was 152.5 and 179.5 ng/mL (p = 0.87) and the median HIC was 9.9 and 8.2 µmol/g dry tissue (p = 0.51), for SVR and NR patients, respectively. Thus, hepatic iron concentration, determined by a reliable quantitative method, was not a negative predictive factor of SVR in patients with chronic hepatitis C presenting mild to moderate hepatic iron accumulation.

Massive Open Online Course management: learning science and engineering through peer-reviewed projects

Massive Open Online Courses (MOOC) are gaining prominence in transversal teaching-learning strategies. However, there are many issues still debated, namely assessment, recognized largely as a cornerstone in Education. The large number of students involved requires a redefinition of strategies that often use approaches based on tasks or challenging projects. In these conditions and due to this approach, assessment is made through peer-reviewed assignments and quizzes online. The peer-reviewed assignments are often based upon sample answers or topics, which guide the student in the task of evaluating peers. This chapter analyzes the grading and evaluation in MOOCs, especially in science and engineering courses, within the context of education and grading methodologies and discusses possible perspectives to pursue grading quality in massive e-learning courses.

Teaching mathematics using Massive Open Online Courses

Distance learning - where students take courses (attend classes, get activities and other sort of learning materials) while being physically separated from their instructors, for larger part of the course duration - is far from being a “new event”. Since the middle of the nineteenth century, this has been done through Radio, Mail and TV, taking advantage of the full educational potential that these media resources had to offer at the time. However, in recent times we have, at our complete disposal, the “magic wonder” of communication and globalization - the Internet. Taking advantage of a whole new set of educational opportunities, with a more or less unselfish “look” to economic interests, focusing its concern on a larger and collective “welfare”, contributing to the development of a more “equitable” world, with regard to educational opportunities, the Massive Open Online Courses (MOOCs) were born and have become an important feature of the higher education in recent years. Many people have been talking about MOOCs as a potential educational revolution, which has arrived from North America, still growing and spreading, referring to its benefits and/or disadvantages. The Polytechnic Institute of Porto, also known as IPP, is a Higher Education Portuguese institution providing undergraduate and graduate studies, which has a solid history of online education and innovation through the use of technology, and it has been particularly interested and focused on MOOC developments, based on an open educational policy in order to try to implement some differentiated learning strategies to its actual students and as a way to attract future ones. Therefore, in July 2014, IPP launched the first Math MOOC on its own platform. This paper describes the requirements, the resulting design and implementation of a mathematics MOOC, which was essentially addressed to three target populations: - pre-college students or individuals wishing to update their Math skills or that need to prepare for the National Exam of Mathematics; - Higher Education students who have not attended in High School, this subject, and who feel the need to acquire basic knowledge about some of the topics covered; - High School Teachers who may use these resources with their students allowing them to develop teaching methodologies like "Flipped Classroom” (available at http://www.opened.ipp.pt/). The MOOC was developed in partnership with several professors from several schools from IPP, gathering different math competences and backgrounds to create and put to work different activities such video lectures and quizzes. We will also try to briefly discuss the advertising strategy being developed to promote this MOOC, since it is not offered through a main MOOC portal, such as Coursera or Udacity.

gvSIG An Open Source GIS solution for the Android Platform

Dissertation submitted in partial fulfilment of the requirements for the Degree of Master of Science in Geospatial Technologies

High-Content Analysis of Breast Cancer Using Single-Cell Deep Transfer Learning

High-content analysis has revolutionized cancer drug discovery by identifying substances that alter the phenotype of a cell, which prevents tumor growth and metastasis. The high-resolution biofluorescence images from assays allow precise quantitative measures enabling the distinction of small molecules of a host cell from a tumor. In this work, we are particularly interested in the application of deep neural networks (DNNs), a cutting-edge machine learning method, to the classification of compounds in chemical mechanisms of action (MOAs). Compound classification has been performed using image-based profiling methods sometimes combined with feature reduction methods such as principal component analysis or factor analysis. In this article, we map the input features of each cell to a particular MOA class without using any treatment-level profiles or feature reduction methods. To the best of our knowledge, this is the first application of DNN in this domain, leveraging single-cell information. Furthermore, we use deep transfer learning (DTL) to alleviate the intensive and computational demanding effort of searching the huge parameter's space of a DNN. Results show that using this approach, we obtain a 30% speedup and a 2% accuracy improvement.