Ramos J.,University of Minho |
Anacleto R.,GECAD |
Costa A.,University of Minho |
Novais P.,University of Minho |
And 2 more authors.
Advances in Intelligent and Soft Computing | Year: 2012
In health care there has been a growing interest and investment in new tools to have a constant monitoring of patients. The increasing of average life expectation and, consequently, the costs in health care due to elderly population are the motivation for this investment.However, health monitoring is not only important to elderly people, it can be also applied to people with cognitive disabilities. In this article we present some systems, which try to support these persons on doing their day-to-day activities and how it can improve their life quality. Also, we present an idea to a project that tries to help the persons with cognitive disabilities by providing assistance in geo-guidance and keep their caregivers aware of their location. © 2012 Springer-Verlag.
Anacleto R.,GECAD |
Luz N.,GECAD |
Almeida A.,GECAD |
Figueiredo L.,GECAD |
Novais P.,University of Minho
Advances in Intelligent and Soft Computing | Year: 2011
Shopping centers present a rich and heterogeneous environment, where IT systems can be implemented in order to support the needs of its actors. However, due to the environment complexity, several feasibility issues emerge when designing both the logical and physical architecture of such systems. Additionally, the system must be able to cope with the individual needs of each actor, and provide services that are easily adopted by them, taking into account several sociological and economical aspects. In this sense, we present an overview of current support systems for shopping center environments. From this overview, a high-level model of the domain (involving actors and services) is described along with challenges and possible features in the context of current Semantic Web, mobile device and sensor technologies. © 2011 Springer-Verlag Berlin Heidelberg.
News Article | November 15, 2016
LONDON, Nov. 15, 2016 /PRNewswire/ -- Gluru (www.gluru.co) today announced the next generation of its Artificial Intelligence-powered task platform with the release of the world's first truly smart task manager -- one of many tools expected to leverage Gluru's machine-learning technology over the next 12 months to supercharge both personal and professional productivity. Gluru is also announcing a $2 million seed round, its second in as many years, bringing the company's total funding to $3.5 million. The latest round includes investment from British venture capital firms Sussex Place Ventures, SAATCHiNVEST, GECAD, and Playfair Capital. Gluru's smart to-do list app, available on web and Android, and in beta on iOS, utilizes a multi-patented unique approach to natural language processing, deep learning, and ontology, syncing with users' conversational data sources, such as Google mail and calendars, to identify important daily tasks and suggest relevant actions. This generates an intelligent "stream" of suggested tasks users most need to accomplish in a day. Gluru continues to learn as users interact with it, automating repetitive items so they can more easily zero in on priority tasks. Users can also schedule manual tasks to speed up Gluru's learning. Soon, Gluru will also suggest actions and answers based on individuals' or companies' entire knowledge bases or service based applications. "Working on the cutting edge of artificial intelligence over the last two years, we have been honing this sophisticated technology to deliver a platform that can empower productivity like never before," said Tim Porter, Founder and CEO of Gluru. "Our smart task manager is only the beginning; our vision is to power consumers' and businesses' productivity platforms, services, and devices with AI, so users can focus on what really matters, saving time and money. Having strategic funding partners like Sussex Place Ventures and SAATCHiNVEST has been paramount to making that a reality." With the latest round of funding, Gluru is positioning its technology to be at the center of productivity, with planned integrations that can essentially add a layer of task and knowledge retrieval through AI-driven deep learning to any platform or tool to make it smarter and more useful. Upcoming integrations include Slack chatbots, Google Drive, Microsoft Outlook and Office, Amazon Alexa, and more. Tiered subscription options for enterprises will also be introduced later this year. "We love groundbreaking enterprise AI that provides practical applications right now, as Gluru is doing – not some distant R&D pipe dream. Tim and the Gluru team have some of the best AI talent in the world, disrupting the productivity space today, and we're excited about where they can take this," said Alex Dunsdon, equity partner at SAATCHiNVEST. For more information about Gluru's smart to-do list, including a demo or screenshots, please contact Elizabeth@ResoundMarketing.com. About Gluru Gluru offers an AI-powered smart task platform that supercharges personal and professional productivity. Gluru's smart to-do list app connects with your email and calendar data to identify the tasks important to you and suggest relevant actions and content, from individual or company-wide knowledge bases or service based applications, to help you achieve your goals, learning as it goes to automatically prioritize your day. The Gluru platform also links with other productivity platforms to make them smarter and more useful, thanks to a highly sophisticated back-end that leverages artificial intelligence, machine learning, and contextual information processing. The Gluru team hails from Google, Apple, Shazam, and Microsoft Swiftkey, with PhDs in machine learning, neural networks, ontologies, and natural language processing. For more ways to organize important moments, visit www.gluru.co. To access the IOS Beta simply go to gluru.co and click on the IOS icon then register.
Chen N.,GECAD |
Ribeiro B.,University of Coimbra |
Vieira A.S.,GECAD |
Duarte J.,GECAD |
Neves J.C.,University of Lisbon
ICMLC 2010 - The 2nd International Conference on Machine Learning and Computing | Year: 2010
Cost-sensitive classification algorithms that enable effective prediction, where the costs of misclassification can be very different, are crucial to creditors and auditors in credit risk analysis. Learning vector quantization (LVQ) is a powerful tool to solve bankruptcy prediction problem as a classification task. The genetic algorithm (GA) is applied widely in conjunction with artificial intelligent methods. The hybridization of genetic algorithm with existing classification algorithms is well illustrated in the field of bankruptcy prediction. In this paper, a hybrid GA and LVQ approach is proposed to minimize the expected misclassified cost under the asymmetric cost preference. Experiments on real-life French private company data show the proposed approach helps to improve the predictive performance in asymmetric cost setup. © 2010 IEEE.
Bras L.,INESC Porto |
Jorge A.M.,INESC Porto |
Gomes E.F.,GECAD |
Duarte R.,Centro Diagnostico Pneumologico Chest Disease Center Vn Gaia
Technology and Medical Sciences, TMSi 2010 - Proceedings of the 6th International Conference on Technology and Medical Sciences | Year: 2011
We are developing a new method for the identification of rib boundaries in chest x-ray images. The identification of rib boundaries is important for radiologist diagnosis of lung diseases as TB. The radiologists use the ribs as reference for location and can be used to eliminate false positives in the detection of abnormalities. Our method automatically identifies rib boundaries from raw images through a sequence of steps using a combination of image processing techniques. Radiographs are still very relevant in practice because in Portugal and many other countries it is the first step for TB detection. We have access a large database of x-ray images provided by the pneumological screening centre (CDP) of Vila Nova de Gaia, in Portugal. © 2011 Taylor & Francis Group.
Ribeiro B.,University of Coimbra |
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2014
Hashing techniques have recently become the trend for accessing complex content over large data sets. With the overwhelming financial data produced today, binary embeddings are efficient tools of indexing big datasets for financial credit risk analysis. The rationale is to find a good hash function such that similar data points in Euclidean space preserve their similarities in the Hamming space for fast data retrieval. In this paper, first we use a semi-supervised hashing method to take into account the pairwise supervised information for constructing the weight adjacency graph matrix needed to learn the binarised Laplacian EigenMap. Second, we train a generalised regression neural network (GRNN) to learn the k-bits hash code. Third, the k-bits code for the test data is efficiently found in the recall phase. The results of hashing financial data show the applicability and advantages of the approach to credit risk assessment. © Springer International Publishing Switzerland 2014.
Chen N.,GECAD |
Ribeiro B.,University of Coimbra |
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2011
As one of the major business problems, corporate bankruptcy has been extensively studied using a large variety of statistical and machine learning approaches. However, the trajectory of bankruptcy behavior is seldom explored in the literature. In this paper, we use self-organizing map neural networks to analyze the changes of financial situation of companies in several consecutive years through a two-step clustering process. Firstly, the bankruptcy risk is characterized by a feature map, and therefore the temporal sequence is converted to the trajectory vector projected on the map. Afterwards, the trajectory map clusters the trajectory vectors to a number of evolution patterns. The approach is applied to a large database of French companies which contains the financial ratios spawning over a period of four years. Typical behaviors such as the deterioration and amelioration associated with the bankruptcy risk, as well as the influence of financial ratios can be revealed by means of visual interpretation. © 2011 Springer-Verlag.
Chen N.,GECAD |
Ribeiro B.,University of Coimbra
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013
Bankruptcy prediction is an extremely important topic in the field of financial decision making. There has been a raising interest in studying more accurate predictive models able to provide valuable early warning before the real business failure. Recent researches suggested using the consensus of multiple classifiers for boosting the prediction performance. Yet rarely the cost of misclassification errors is considered in the literature of consensus decision making. In this paper we investigate the performance of classifier ensembles for cost-sensitive bankruptcy prediction. The selection of ensemble members is based on individual performance and pairwise diversity of classifiers. The experimental results on a real world database of French companies show that by selecting appropriate base classifiers the ensemble learning substantially improves the performance of cost-sensitive bankruptcy prediction. © 2013 Springer-Verlag Berlin Heidelberg.
Pereira J.,UTAD |
Costa C.,UTAD |
Silva D.,UTAD |
Varajao J.,ALGORITMI |
9th European Conference on eLearning 2010, ECEL 2010 | Year: 2010
While much information is available on pedagogic uses of virtual worlds, with Second Life being the most common virtual world platform in current educational literature, an organization must consider its presence in this environment as more than the mere sum of individual educational efforts. Resources need to be shared between educational stakeholders, visual navigation needs to make sense, and the sense of being within an actual organization should be conveyed (not just the sense of being within a collection of personal spaces). But there is little information on how a virtual campus for an educational organization should be structured. Virtual campi in Second Life for adult education institutions don't typically reproduce their physical counterparts. While spaces such as lecture halls, amphitheatres, meeting places, and libraries are commonly found, the specific features of the medium imply an organization of spaces and usage that differ from physical campi. For instance, navigational affordances are different (ability to fly and gravity-immune objects, for instance), as are communicational features (specific limits on the reach of voice and text communication), and user involvement (how students and teachers use the spaces). We conducted a survey of several existing Second Life campi of adult education institutions (mostly universities), to establish what spaces are present in each and how they are used and organized. In this paper, we present the overall process, and the structure and instructions for data collection by all people involved. Then we detail the various kinds of spaces (by function, not by aesthetic) found in the campi and their prevalence. We also present data on user-oriented features of the campi, and crossanalyse this with their occurrence per space and campi. This survey was part of the process for specification and development of the virtual Second Life campus for project VITA, a EC-funded project to create and experiment learning actions directed to SME' managers for development of entrepreneurship competences. Thus, we conclude with an example of how the survey results can be used to support the development of campi, by briefly presenting the campus that was developed specifically for this project.
Bettencourt N.,GECAD |
Silva N.,GECAD |
Barroso J.,University of Trás os Montes e Alto Douro
CEUR Workshop Proceedings | Year: 2014
In a world overwhelmed by constant data creation and manipulation, where privacy is becoming a real concern, topics like data usage control, accountability, provenance, protected sharing of resources and trustworthiness of knowledge sources are becoming main topics of discussion among communities of interest. In this paper enhancements are proposed for an existing framework that tackles some of the afore mentioned issues namely data provenance, usage control and accountability. Such proposals consist of providing means for publishing resources in a private manner hereby making websites behave like meshes of hyperlinked resources from different domains, not only for resources publicly published but also for the ones protected by access policies. © 2014, Society, Privacy and the Semantic Web Policy and Technology.