Publications

Aliasing control with view-based typestate

Militao F., Aldrich J., Caires L.

24th European Conference on Object-Oriented Programming, ECOOP 2010 Workshop Proceedings - Workshop 5:12th Workshop on Formal Techniques for Java-Like Programs, FTFJP'10

2010

Conference Paper

Abstract:

Tracking the state of an object (in the sense of how a File can be in an Open or Closed state) is difficult not just because of the problem of managing state transitions but also due to the complexity introduced by aliasing. Unchecked duplication of object references makes local reasoning impossible by allowing situations where transitions can be triggered unexpectedly (for instance, passing aliased parameters to a method that expects unaliased parameters, or calling a method that has a side effect through an alias deeply nested in a data structure). We propose a generalization of access permissions that goes beyond a fixed set of permissions to an object. In this paper we present a new aliasing control mechanism that uses a small set of permissions as building block for the creation of views that capture a projection of an object with specific access constraints to its fields and/or methods. This makes permission tracking more fine grained while also making the designer’s intent more explicit. We present a few meaningful examples of how these views handle situations such as: separating different sections of an object for safe initialization; and access with either an unbounded number of readers or a single writer (multiple readers or unique writer). Finally, we show a type system for checking correctness of state use in the presence of this kind of controlled aliasing.

Alternating direction optimization for image segmentation using hidden Markov measure field models

Bioucas-Dias J., Condessa F., Kovacevic J.

Proceedings of SPIE - The International Society for Optical Engineering

2014

Conference Paper

Abstract:

Image segmentation is fundamentally a discrete problem. It consists of finding a partition of the image domain such that the pixels in each element of the partition exhibit some kind of similarity. The solution is often obtained by minimizing an objective function containing terms measuring the consistency of the candidate partition with respect to the observed image, and regularization terms promoting solutions with desired properties. This formulation ends up being an integer optimization problem that, apart from a few exceptions, is NP-hard and thus impossible to solve exactly. This roadblock has stimulated active research aimed at computing “good” approximations to the solutions of those integer optimization problems. Relevant lines of attack have focused on the representation of the regions (i.e., the partition elements) in terms of functions, instead of subsets, and on convex relaxations which can be solved in polynomial time. In this paper, inspired by the “hidden Markov measure field” introduced by Marroquin et al. in 2003, we sidestep the discrete nature of image segmentation by formulating the problem in the Bayesian framework and introducing a hidden set of real-valued random fields determining the probability of a given partition. Armed with this model, the original discrete optimization is converted into a convex program. To infer the hidden fields, we introduce the Segmentation via the Constrained Split Augmented Lagrangian Shrinkage Algorithm (SegSALSA). The effectiveness of the proposed methodology is illustrated with simulated and real hyperspectral and medical images.

An Analytical Model for Efficient Indoor THz Access Point Deployment

Singh R., Sicker D.

IEEE Wireless Communications and Networking Conference

2020

Conference Paper

Abstract:

Ultra-densification of user equipment (UE) and access points (APs) are anticipated to take a toll on the future spectrum needs. Higher frequency bands, such as mmWave (30-300GHz) and THz spectrum (0.3-10THz), can be used to cater to the high-throughput needs of ultra-dense networks. These high-frequency bands have a tremendous amount ofgreen-filed contiguous spectrum, ranging in hundreds of GHz. However, these bands, especially the THz bands, face nu- merous challenges, such as high spreading, absorption, and penetration losses. To combat these challenges, the THz-APs need to be either equipped with high transmit power, high antenna gains (i.e., narrow antenna beams), or limit the com- munication to short-ranges. All of these factors are bounded due to technical or economic challenges, which will result in a“distance-power dilemma” while deciding on the deployment strategy of THz-APs. In this paper, we present an analytical model to deploy THz-APs in an indoor setting efficiently. We further show through extensive numerical analysis, the optimal number of APs and optimal room length for different blocks of the THz spectrum. Furthermore, these THz-APs need to be efficiently packed to avoid outages due to handoffs, which can add more complexity to the dilemma. To mitigate the packing problem, we propose two solutions over the optimal solution: (a) Radius Increase, and (b) Repeater Assistance, and present an analytical model for each.

An Approach for Personalized Social Interactions Between a Therapeutic Robot and Children with Autism Spectrum Disorder

Baraka K., Couto M., Melo F.S., Paiva A., Veloso M.

GAIPS Technical Report Series

2019

Other

Abstract:

For high-variability populations, such as individuals with Autism Spectrum Disorders (ASD), the importance of personalization and adaptation mechanisms in Human-Robot Interaction becomes crucial. This technical report presents an algorithmic method for personalization of robotic behavior in structured social interactions with ASD children, based on state-of-the-art diagnostic models. In a first step, we leverage the structure of the diagnostic procedure to build robotic behaviors on a NAO humanoid robot, aimed at eliciting target behaviors from the child. Through appropriate sequencing of possible actions, the robot is able to assess a child’s behavioral profile and use it to personalize the interaction. To test our method, we developed a semi-autonomous robotic scenario where a humanoid robot interacts with a child with ASD through interactive storytelling, focusing on social prompts related to deficits in attention, one of the core impairments of ASD. We present the design and methodology of an evaluation study run with 11 young ASD children in a child development center in Portugal.

An augmented Lagrangian approach to constrained MAP inference

Martins A.F.T., Figueiredo M.A.T., Aguiar P.M.Q., Smith N.A., Xing E.P.

Proceedings of the 28th International Conference on Machine Learning, ICML 2011

2011

Conference Paper

Abstract:

We propose a new algorithm for approximate MAP inference on factor graphs, by combining augmented Lagrangian optimization with the dual decomposition method. Each slave subproblem is given a quadratic penalty, which pushes toward faster consensus than in previous subgradient approaches. Our algorithm is provably convergent, parallelizable, and suitable for fine decompositions of the graph. We show how it can efficiently handle problems with (possibly global) structural constraints via simple sort operations. Experiments on synthetic and real-world data show that our approach compares favorably with the state-of-the-art.

An automatic subject-Adaptable heartbeat classifier based on multi-view learning

Ye C., Vijaya Kumar B.V.K., Coimbra M.T.

IEEE Journal of Biomedical and Health Informatics

2015

Article

Abstract:

In this paper, a novel subject-adaptable heartbeat classification model is presented, in order to address the significant interperson variations in ECG signals. A multiview learning approach is proposed to automate subject adaptation using a small amount of unlabeled personal data, without requiring manual labeling. The designed subject-customized models consist of two models, namely, general classification model and specific classification model. The general model is trained using similar subjects out of a population dataset, where a pattern matching based algorithm is developed to select the subjects that are “similar” to the particular test subject for model training. In contrast, the specific model is trained mainly on a small amount of high-confidence personal dataset, resulting from multiview-based learning. The learned general model represents the population knowledge, providing an interperson perspective for classification, while the specific model corresponds to the specific knowledge of the subject, offering an intraperson perspective for classification. The two models supplement each other and are combined to achieve improved personalized ECG analysis. The proposed methods have been validated on the MIT-BIH Arrhythmia Database, yielding an average classification accuracy of 99.4% for ventricular ectopic beat class and 98.3% for supraventricular ectopic beat class, which corresponds to a significant improvement over other published results.

An Availability Model for DSS and OLTP Applications in Virtualized Environments

Torquato M., Gonçalves C.F., Vieira M.

16th European Dependable Computing Conference (EDCC 2020)

2020

Conference Paper

Abstract:

Decision support systems (DSS) and online transaction processing applications (OLTP) are crucial for several organizations and frequently require high levels of availability. Many organizations moved their systems to the virtualized environment aiming at improving system availability. Despite the flexibility and manageability features provided by virtualization, a question arises on what policies to apply in order to achieve high availability. Usual approaches highlight redundancy as a strategy for high availability. Still, a concern persists on what components we should consider for redundancy. This paper proposes a hierarchical availability model for evaluating different redundancy allocations for DSS and OLTP systems in virtualized environments. We present three case studies investigating only-Virtual Machine (VM) redundancy and physical machine redundancy strategies. The results provide an overview of the availability impact due to each strategy. We noticed that the physical machine failure rate limits the maximum availability obtained from only-VM redundancy. We exercise our model with a genetic algorithm to find alternatives for high availability. The presented models and results may bring insights when designing availability policies.

An effective method for modeling wind power forecast uncertainty

Mauch B., Apt J., Carvalho P.M.S., Small M.J.

Energy Systems

2013

Article

Abstract:

Wind forecasts are an important tool for electric system operators. Proper use of wind power forecasts to make operating decisions must account for the uncertainty associated with the forecast. Data from different regions in the USA with forecasts made by different vendors show the forecast error distribution is strongly dependent on the forecast level of wind power. For low wind power forecast, the forecasts tend to under-predict the actual wind power produced, whereas when the forecast is for high power, the forecast tends to over-predict the actual wind power. Most of the work in this field neglects the influence of wind forecast levels on wind forecast uncertainty and analyzes wind forecast errors as a whole. The few papers that account for this dependence, group wind forecasts by the value of the forecast and fit parametric distributions to actual wind power in each bin of data. In the latter case, different parameters and possibly different distributions are estimated for each data bin. We present a method to model wind power forecast uncertainty as a single closed-form solution using a logit transformation of historical wind power forecast and actual wind power data. Once transformed, the data become close to jointly normally distributed. We show the process of calculating confidence intervals of wind power forecast errors using the jointly normally distributed logit transformed data. This method has the advantage of fitting the entire dataset with five parameters while also providing the ability to make calculations conditioned on the value of the wind power forecast.

An IEEE 802.11p/WAVE implementation with synchronous channel switching for seamless dual-channel access (poster)

Ameixieira C., Matos J., Moreira R., Cardote A., Oliveira A., Sargento S.

IEEE Vehicular Networking Conference, VNC

2011

Conference Paper

Abstract:

The emerging interest in vehicular networks (VANET) led to the deployment of several small and medium-scale testbeds to evaluate the characteristics of this technology in real-world scenarios. Despite this, due to the low availability and high cost of IEEE 802.11p/WAVE fully compliant hardware and software, many of these experiments have been performed with other communication standards, which generate, in many cases, misleading results, that are not representative of real-world vehicular communications. As a mean of solving this problem, this paper presents the implementation and validation of a fully compliant MAC/PHY solution developed in the scope of the DRIVE-IN project, which will be used in a 500-node testbed. Contrarily to what happens with most of the solutions existent in the market, our system complies with the strict channel switching timings using GPS time synchronization, providing access to two different types of wireless channels (e.g. control and service channels as defined in IEEE 1609.4) in a seamless way for the end user. This feature allows safety-critical and control messages to be sent in a dedicated channel, with very low latency, while another channel may be used for less critical services (e.g. infotainment applications, advertisement). The results of the conducted tests show that indeed it is possible to assure the support of critical safety services in the presence of other traffic.