Showing 1–42 of 42 results for author: Pinto, R

3W Dataset 2.0.0: a realistic and public dataset with rare undesirable real events in oil wells

Authors: Ricardo Emanuel Vaz Vargas, Afrânio José de Melo Junior, Celso José Munaro, Cláudio Benevenuto de Campos Lima, Eduardo Toledo de Lima Junior, Felipe Muntzberg Barrocas, Flávio Miguel Varejão, Guilherme Fidelis Peixer, Igor de Melo Nery Oliveira, Jader Riso Barbosa Jr., Jaime Andrés Lozano Cadena, Jean Carlos Dias de Araújo, João Neuenschwander Escosteguy Carneiro, Lucas Gouveia Omena Lopes, Lucas Pereira de Gouveia, Mateus de Araujo Fernandes, Matheus Lima Scramignon, Patrick Marques Ciarelli, Rodrigo Castello Branco, Rogério Leite Alves Pinto

Abstract: In the oil industry, undesirable events in oil wells can cause economic losses, environmental accidents, and human casualties. Solutions based on Artificial Intelligence and Machine Learning for Early Detection of such events have proven valuable for diverse applications across industries. In 2019, recognizing the importance and the lack of public datasets related to undesirable events in oil well… ▽ More In the oil industry, undesirable events in oil wells can cause economic losses, environmental accidents, and human casualties. Solutions based on Artificial Intelligence and Machine Learning for Early Detection of such events have proven valuable for diverse applications across industries. In 2019, recognizing the importance and the lack of public datasets related to undesirable events in oil wells, Petrobras developed and publicly released the first version of the 3W Dataset, which is essentially a set of Multivariate Time Series labeled by experts. Since then, the 3W Dataset has been developed collaboratively and has become a foundational reference for numerous works in the field. This data article describes the current publicly available version of the 3W Dataset, which contains structural modifications and additional labeled data. The detailed description provided encourages and supports the 3W community and new 3W users to improve previous published results and to develop new robust methodologies, digital products and services capable of detecting undesirable events in oil wells with enough anticipation to enable corrective or mitigating actions. △ Less

Submitted 25 June, 2025; originally announced July 2025.

Comments: 21 pages, 10 figures, and 7 tables

Neuroevolution of Self-Attention Over Proto-Objects

Authors: Rafael C. Pinto, Anderson R. Tavares

Abstract: Proto-objects - image regions that share common visual properties - offer a promising alternative to traditional attention mechanisms based on rectangular-shaped image patches in neural networks. Although previous work demonstrated that evolving a patch-based hard-attention module alongside a controller network could achieve state-of-the-art performance in visual reinforcement learning tasks, our… ▽ More Proto-objects - image regions that share common visual properties - offer a promising alternative to traditional attention mechanisms based on rectangular-shaped image patches in neural networks. Although previous work demonstrated that evolving a patch-based hard-attention module alongside a controller network could achieve state-of-the-art performance in visual reinforcement learning tasks, our approach leverages image segmentation to work with higher-level features. By operating on proto-objects rather than fixed patches, we significantly reduce the representational complexity: each image decomposes into fewer proto-objects than regular patches, and each proto-object can be efficiently encoded as a compact feature vector. This enables a substantially smaller self-attention module that processes richer semantic information. Our experiments demonstrate that this proto-object-based approach matches or exceeds the state-of-the-art performance of patch-based implementations with 62% less parameters and 2.6 times less training time. △ Less

Submitted 30 April, 2025; originally announced May 2025.

Comments: 9 pages, 16 figures, GECCO

A new method for erasure decoding of convolutional codes

Authors: Julia Lieb, Raquel Pinto, Carlos Vela

Abstract: In this paper, we propose a new erasure decoding algorithm for convolutional codes using the generator matrix. This implies that our decoding method also applies to catastrophic convolutional codes in opposite to the classic approach using the parity-check matrix. We compare the performance of both decoding algorithms. Moreover, we enlarge the family of optimal convolutional codes (complete-MDP) b… ▽ More In this paper, we propose a new erasure decoding algorithm for convolutional codes using the generator matrix. This implies that our decoding method also applies to catastrophic convolutional codes in opposite to the classic approach using the parity-check matrix. We compare the performance of both decoding algorithms. Moreover, we enlarge the family of optimal convolutional codes (complete-MDP) based on the generator matrix. △ Less

Submitted 22 April, 2025; originally announced April 2025.

Enhancing Diffusion Models for High-Quality Image Generation

Authors: Jaineet Shah, Michael Gromis, Rickston Pinto

Abstract: This report presents the comprehensive implementation, evaluation, and optimization of Denoising Diffusion Probabilistic Models (DDPMs) and Denoising Diffusion Implicit Models (DDIMs), which are state-of-the-art generative models. During inference, these models take random noise as input and iteratively generate high-quality images as output. The study focuses on enhancing their generative capabil… ▽ More This report presents the comprehensive implementation, evaluation, and optimization of Denoising Diffusion Probabilistic Models (DDPMs) and Denoising Diffusion Implicit Models (DDIMs), which are state-of-the-art generative models. During inference, these models take random noise as input and iteratively generate high-quality images as output. The study focuses on enhancing their generative capabilities by incorporating advanced techniques such as Classifier-Free Guidance (CFG), Latent Diffusion Models with Variational Autoencoders (VAE), and alternative noise scheduling strategies. The motivation behind this work is the growing demand for efficient and scalable generative AI models that can produce realistic images across diverse datasets, addressing challenges in applications such as art creation, image synthesis, and data augmentation. Evaluations were conducted on datasets including CIFAR-10 and ImageNet-100, with a focus on improving inference speed, computational efficiency, and image quality metrics like Frechet Inception Distance (FID). Results demonstrate that DDIM + CFG achieves faster inference and superior image quality. Challenges with VAE and noise scheduling are also highlighted, suggesting opportunities for future optimization. This work lays the groundwork for developing scalable, efficient, and high-quality generative AI systems to benefit industries ranging from entertainment to robotics. △ Less

Submitted 18 December, 2024; originally announced December 2024.

Testing Support Size More Efficiently Than Learning Histograms

Authors: Renato Ferreira Pinto Jr., Nathaniel Harms

Abstract: Consider two problems about an unknown probability distribution $p$: 1. How many samples from $p$ are required to test if $p$ is supported on $n$ elements or not? Specifically, given samples from $p$, determine whether it is supported on at most $n$ elements, or it is "$ε$-far" (in total variation distance) from being supported on $n$ elements. 2. Given $m$ samples from $p$, what is the larges… ▽ More Consider two problems about an unknown probability distribution $p$: 1. How many samples from $p$ are required to test if $p$ is supported on $n$ elements or not? Specifically, given samples from $p$, determine whether it is supported on at most $n$ elements, or it is "$ε$-far" (in total variation distance) from being supported on $n$ elements. 2. Given $m$ samples from $p$, what is the largest lower bound on its support size that we can produce? The best known upper bound for problem (1) uses a general algorithm for learning the histogram of the distribution $p$, which requires $Θ(\tfrac{n}{ε^2 \log n})$ samples. We show that testing can be done more efficiently than learning the histogram, using only $O(\tfrac{n}{ε\log n} \log(1/ε))$ samples, nearly matching the best known lower bound of $Ω(\tfrac{n}{ε\log n})$. This algorithm also provides a better solution to problem (2), producing larger lower bounds on support size than what follows from previous work. The proof relies on an analysis of Chebyshev polynomial approximations outside the range where they are designed to be good approximations, and the paper is intended as an accessible self-contained exposition of the Chebyshev polynomial method. △ Less

Submitted 31 March, 2025; v1 submitted 24 October, 2024; originally announced October 2024.

Comments: 40 pages. Minor edits, added Open questions

PReLU: Yet Another Single-Layer Solution to the XOR Problem

Authors: Rafael C. Pinto, Anderson R. Tavares

Abstract: This paper demonstrates that a single-layer neural network using Parametric Rectified Linear Unit (PReLU) activation can solve the XOR problem, a simple fact that has been overlooked so far. We compare this solution to the multi-layer perceptron (MLP) and the Growing Cosine Unit (GCU) activation function and explain why PReLU enables this capability. Our results show that the single-layer PReLU ne… ▽ More This paper demonstrates that a single-layer neural network using Parametric Rectified Linear Unit (PReLU) activation can solve the XOR problem, a simple fact that has been overlooked so far. We compare this solution to the multi-layer perceptron (MLP) and the Growing Cosine Unit (GCU) activation function and explain why PReLU enables this capability. Our results show that the single-layer PReLU network can achieve 100\% success rate in a wider range of learning rates while using only three learnable parameters. △ Less

Submitted 16 September, 2024; originally announced September 2024.

Directed Isoperimetry and Monotonicity Testing: A Dynamical Approach

Authors: Renato Ferreira Pinto Jr

Abstract: This paper explores the connection between classical isoperimetric inequalities, their directed analogues, and monotonicity testing. We study the setting of real-valued functions $f : [0,1]^d \to \mathbb{R}$ on the solid unit cube, where the goal is to test with respect to the $L^p$ distance. Our goals are twofold: to further understand the relationship between classical and directed isoperimetry,… ▽ More This paper explores the connection between classical isoperimetric inequalities, their directed analogues, and monotonicity testing. We study the setting of real-valued functions $f : [0,1]^d \to \mathbb{R}$ on the solid unit cube, where the goal is to test with respect to the $L^p$ distance. Our goals are twofold: to further understand the relationship between classical and directed isoperimetry, and to give a monotonicity tester with sublinear query complexity in this setting. Our main results are 1) an $L^2$ monotonicity tester for $M$-Lipschitz functions with query complexity $\widetilde O(\sqrt{d} M^2 / ε^2)$ and, behind this result, 2) the directed Poincaré inequality $\mathsf{dist}^{\mathsf{mono}}_2(f)^2 \le C \mathbb{E}[|\nabla^- f|^2]$, where the "directed gradient" operator $\nabla^-$ measures the local violations of monotonicity of $f$. To prove the second result, we introduce a partial differential equation (PDE), the directed heat equation, which takes a one-dimensional function $f$ into a monotone function $f^*$ over time and enjoys many desirable analytic properties. We obtain the directed Poincaré inequality by combining convergence aspects of this PDE with the theory of optimal transport. Crucially for our conceptual motivation, this proof is in complete analogy with the mathematical physics perspective on the classical Poincaré inequality, namely as characterizing the convergence of the standard heat equation toward equilibrium. △ Less

Submitted 1 October, 2024; v1 submitted 27 April, 2024; originally announced April 2024.

Comments: 86 pages; added comments to improve the readability of the paper, and small edits to the intro

Distribution Testing with a Confused Collector

Authors: Renato Ferreira Pinto Jr., Nathaniel Harms

Abstract: We are interested in testing properties of distributions with systematically mislabeled samples. Our goal is to make decisions about unknown probability distributions, using a sample that has been collected by a confused collector, such as a machine-learning classifier that has not learned to distinguish all elements of the domain. The confused collector holds an unknown clustering of the domain a… ▽ More We are interested in testing properties of distributions with systematically mislabeled samples. Our goal is to make decisions about unknown probability distributions, using a sample that has been collected by a confused collector, such as a machine-learning classifier that has not learned to distinguish all elements of the domain. The confused collector holds an unknown clustering of the domain and an input distribution $μ$, and provides two oracles: a sample oracle which produces a sample from $μ$ that has been labeled according to the clustering; and a label-query oracle which returns the label of a query point $x$ according to the clustering. Our first set of results shows that identity, uniformity, and equivalence of distributions can be tested efficiently, under the earth-mover distance, with remarkably weak conditions on the confused collector, even when the unknown clustering is adversarial. This requires defining a variant of the distribution testing task (inspired by the recent testable learning framework of Rubinfeld & Vasilyan), where the algorithm should test a joint property of the distribution and its clustering. As an example, we get efficient testers when the distribution tester is allowed to reject if it detects that the confused collector clustering is "far" from being a decision tree. The second set of results shows that we can sometimes do significantly better when the clustering is random instead of adversarial. For certain one-dimensional random clusterings, we show that uniformity can be tested under the TV distance using $\widetilde O\left(\frac{\sqrt n}{ρ^{3/2} ε^2}\right)$ samples and zero queries, where $ρ\in (0,1]$ controls the "resolution" of the clustering. We improve this to $O\left(\frac{\sqrt n}{ρε^2}\right)$ when queries are allowed. △ Less

Submitted 23 November, 2023; originally announced November 2023.

Comments: 64 pages. Full version of paper to appear at ITCS 2024. arXiv admin note: text overlap with arXiv:2304.01374

Directed Poincaré Inequalities and $L^1$ Monotonicity Testing of Lipschitz Functions

Authors: Renato Ferreira Pinto Jr

Abstract: We study the connection between directed isoperimetric inequalities and monotonicity testing. In recent years, this connection has unlocked breakthroughs for testing monotonicity of functions defined on discrete domains. Inspired the rich history of isoperimetric inequalities in continuous settings, we propose that studying the relationship between directed isoperimetry and monotonicity in such se… ▽ More We study the connection between directed isoperimetric inequalities and monotonicity testing. In recent years, this connection has unlocked breakthroughs for testing monotonicity of functions defined on discrete domains. Inspired the rich history of isoperimetric inequalities in continuous settings, we propose that studying the relationship between directed isoperimetry and monotonicity in such settings is essential for understanding the full scope of this connection. Hence, we ask whether directed isoperimetric inequalities hold for functions $f : [0,1]^n \to \mathbb{R}$, and whether this question has implications for monotonicity testing. We answer both questions affirmatively. For Lipschitz functions $f : [0,1]^n \to \mathbb{R}$, we show the inequality $d^{\mathsf{mono}}_1(f) \lesssim \mathbb{E}\left[\|\nabla^- f\|_1\right]$, which upper bounds the $L^1$ distance to monotonicity of $f$ by a measure of its "directed gradient". A key ingredient in our proof is the monotone rearrangement of $f$, which generalizes the classical "sorting operator" to continuous settings. We use this inequality to give an $L^1$ monotonicity tester for Lipschitz functions $f : [0,1]^n \to \mathbb{R}$, and this framework also implies similar results for testing real-valued functions on the hypergrid. △ Less

Submitted 5 July, 2023; originally announced July 2023.

Comments: 35 pages including 5 page appendix. To appear at RANDOM 2023

A new construction of an MDS convolutional code of rate 1/2

Authors: Zita Abreu, Raquel Pinto, Rita Simões

Abstract: Maximum distance separable convolutional codes are characterized by the property that the free distance reaches the generalized Singleton bound, which makes them optimal for error correction. However, the existing constructions of such codes are available over fields of large size. In this paper, we present the unique construction of MDS convolutional codes of rate 1/2 and degree 5 over the field… ▽ More Maximum distance separable convolutional codes are characterized by the property that the free distance reaches the generalized Singleton bound, which makes them optimal for error correction. However, the existing constructions of such codes are available over fields of large size. In this paper, we present the unique construction of MDS convolutional codes of rate 1/2 and degree 5 over the field F_{11}. △ Less

Submitted 25 May, 2023; v1 submitted 8 May, 2023; originally announced May 2023.

Criteria for the construction of MDS convolutional codes with good column distances

Authors: Zita Abreu, Julia Lieb, Raquel Pinto, Joachim Rosenthal

Abstract: Maximum-distance separable (MDS) convolutional codes are characterized by the property that their free distance reaches the generalized Singleton bound. In this paper, new criteria to construct MDS convolutional codes are presented. Additionally, the obtained convolutional codes have optimal first (reverse) column distances and the criteria allow to relate the construction of MDS convolutional cod… ▽ More Maximum-distance separable (MDS) convolutional codes are characterized by the property that their free distance reaches the generalized Singleton bound. In this paper, new criteria to construct MDS convolutional codes are presented. Additionally, the obtained convolutional codes have optimal first (reverse) column distances and the criteria allow to relate the construction of MDS convolutional codes to the construction of reverse superregular Toeplitz matrices. Moreover, we present some construction examples for small code parameters over small finite fields. △ Less

Submitted 25 May, 2023; v1 submitted 8 May, 2023; originally announced May 2023.

Distribution Testing Under the Parity Trace

Authors: Renato Ferreira Pinto Jr., Nathaniel Harms

Abstract: Distribution testing is a fundamental statistical task with many applications, but we are interested in a variety of problems where systematic mislabelings of the sample prevent us from applying the existing theory. To apply distribution testing to these problems, we introduce distribution testing under the parity trace, where the algorithm receives an ordered sample $S$ that reveals only the leas… ▽ More Distribution testing is a fundamental statistical task with many applications, but we are interested in a variety of problems where systematic mislabelings of the sample prevent us from applying the existing theory. To apply distribution testing to these problems, we introduce distribution testing under the parity trace, where the algorithm receives an ordered sample $S$ that reveals only the least significant bit of each element. This abstraction reveals connections between the following three problems of interest, allowing new upper and lower bounds: 1. In distribution testing with a confused collector, the collector of the sample may be incapable of distinguishing between nearby elements of a domain (e.g. a machine learning classifier). We prove bounds for distribution testing with a confused collector on domains structured as a cycle or a path. 2. Recent work on the fundamental testing vs. learning question established tight lower bounds on distribution-free sample-based property testing by reduction from distribution testing, but the tightness is limited to symmetric properties. The parity trace allows a broader family of equivalences to non-symmetric properties, while recovering and strengthening many of the previous results with a different technique. 3. We give the first results for property testing in the well-studied trace reconstruction model, where the goal is to test whether an unknown string $x$ satisfies some property or is far from satisfying that property, given only independent random traces of $x$. Our main technical result is a tight bound of $\widetilde Θ\left((n/ε)^{4/5} + \sqrt n/ε^2\right)$ for testing uniformity of distributions over $[n]$ under the parity trace, leading also to results for the problems above. △ Less

Submitted 3 April, 2023; originally announced April 2023.

Comments: 132 pages

Seamless Multimodal Biometrics for Continuous Personalised Wellbeing Monitoring

Authors: João Ribeiro Pinto

Abstract: Artificially intelligent perception is increasingly present in the lives of every one of us. Vehicles are no exception, (...) In the near future, pattern recognition will have an even stronger role in vehicles, as self-driving cars will require automated ways to understand what is happening around (and within) them and act accordingly. (...) This doctoral work focused on advancing in-vehicle sensi… ▽ More Artificially intelligent perception is increasingly present in the lives of every one of us. Vehicles are no exception, (...) In the near future, pattern recognition will have an even stronger role in vehicles, as self-driving cars will require automated ways to understand what is happening around (and within) them and act accordingly. (...) This doctoral work focused on advancing in-vehicle sensing through the research of novel computer vision and pattern recognition methodologies for both biometrics and wellbeing monitoring. The main focus has been on electrocardiogram (ECG) biometrics, a trait well-known for its potential for seamless driver monitoring. Major efforts were devoted to achieving improved performance in identification and identity verification in off-the-person scenarios, well-known for increased noise and variability. Here, end-to-end deep learning ECG biometric solutions were proposed and important topics were addressed such as cross-database and long-term performance, waveform relevance through explainability, and interlead conversion. Face biometrics, a natural complement to the ECG in seamless unconstrained scenarios, was also studied in this work. The open challenges of masked face recognition and interpretability in biometrics were tackled in an effort to evolve towards algorithms that are more transparent, trustworthy, and robust to significant occlusions. Within the topic of wellbeing monitoring, improved solutions to multimodal emotion recognition in groups of people and activity/violence recognition in in-vehicle scenarios were proposed. At last, we also proposed a novel way to learn template security within end-to-end models, dismissing additional separate encryption processes, and a self-supervised learning approach tailored to sequential data, in order to ensure data security and optimal performance. (...) △ Less

Submitted 19 January, 2023; v1 submitted 8 January, 2023; originally announced January 2023.

Comments: Doctoral thesis presented and approved on the 21st of December 2022 to the University of Porto

Android Private Compute Core Architecture

Authors: Eugenio Marchiori, Sarah de Haas, Sergey Volnov, Ronnie Falcon, Roxanne Pinto, Marco Zamarato

Abstract: Android's Private Compute Core (PCC) is a secure, isolated environment within the operating system, that maintains separation from apps while enabling users and developers to maintain control over their data. It is backed by open-source code in the Android Framework introduced in Android 12. PCC allows features to communicate with a server to receive model updates and contribute to global model tr… ▽ More Android's Private Compute Core (PCC) is a secure, isolated environment within the operating system, that maintains separation from apps while enabling users and developers to maintain control over their data. It is backed by open-source code in the Android Framework introduced in Android 12. PCC allows features to communicate with a server to receive model updates and contribute to global model training through Private Compute Services (PCS), the core of which has been open sourced. PCC is part of the OS, and by virtue of being isolated, constrained, and trusted, it can host sophisticated ML features. The hosted features themselves, running inside PCC, can be closed source and updatable. In this way, PCC enables machine learning features to process ambient and OS-level data and improve over time, while restricting the availability of information about individual users to servers or apps. △ Less

Submitted 22 September, 2022; v1 submitted 21 September, 2022; originally announced September 2022.

Causality-Inspired Taxonomy for Explainable Artificial Intelligence

Authors: Pedro C. Neto, Tiago Gonçalves, João Ribeiro Pinto, Wilson Silva, Ana F. Sequeira, Arun Ross, Jaime S. Cardoso

Abstract: As two sides of the same coin, causality and explainable artificial intelligence (xAI) were initially proposed and developed with different goals. However, the latter can only be complete when seen through the lens of the causality framework. As such, we propose a novel causality-inspired framework for xAI that creates an environment for the development of xAI approaches. To show its applicability… ▽ More As two sides of the same coin, causality and explainable artificial intelligence (xAI) were initially proposed and developed with different goals. However, the latter can only be complete when seen through the lens of the causality framework. As such, we propose a novel causality-inspired framework for xAI that creates an environment for the development of xAI approaches. To show its applicability, biometrics was used as case study. For this, we have analysed 81 research papers on a myriad of biometric modalities and different tasks. We have categorised each of these methods according to our novel xAI Ladder and discussed the future directions of the field. △ Less

Submitted 4 March, 2024; v1 submitted 19 August, 2022; originally announced August 2022.

OCFR 2022: Competition on Occluded Face Recognition From Synthetically Generated Structure-Aware Occlusions

Authors: Pedro C. Neto, Fadi Boutros, Joao Ribeiro Pinto, Naser Damer, Ana F. Sequeira, Jaime S. Cardoso, Messaoud Bengherabi, Abderaouf Bousnat, Sana Boucheta, Nesrine Hebbadj, Mustafa Ekrem Erakın, Uğur Demir, Hazım Kemal Ekenel, Pedro Beber de Queiroz Vidal, David Menotti

Abstract: This work summarizes the IJCB Occluded Face Recognition Competition 2022 (IJCB-OCFR-2022) embraced by the 2022 International Joint Conference on Biometrics (IJCB 2022). OCFR-2022 attracted a total of 3 participating teams, from academia. Eventually, six valid submissions were submitted and then evaluated by the organizers. The competition was held to address the challenge of face recognition in th… ▽ More This work summarizes the IJCB Occluded Face Recognition Competition 2022 (IJCB-OCFR-2022) embraced by the 2022 International Joint Conference on Biometrics (IJCB 2022). OCFR-2022 attracted a total of 3 participating teams, from academia. Eventually, six valid submissions were submitted and then evaluated by the organizers. The competition was held to address the challenge of face recognition in the presence of severe face occlusions. The participants were free to use any training data and the testing data was built by the organisers by synthetically occluding parts of the face images using a well-known dataset. The submitted solutions presented innovations and performed very competitively with the considered baseline. A major output of this competition is a challenging, realistic, and diverse, and publicly available occluded face recognition benchmark with well defined evaluation protocols. △ Less

Submitted 15 August, 2022; v1 submitted 4 August, 2022; originally announced August 2022.

Comments: Accepted at International Joint Conference on Biometrics 2022

Predição de Incidência de Lesão por Pressão em Pacientes de UTI usando Aprendizado de Máquina

Authors: Henrique P. Silva, Arthur D. Reys, Daniel S. Severo, Dominique H. Ruther, Flávio A. O. B. Silva, Maria C. S. S. Guimarães, Roberto Z. A. Pinto, Saulo D. S. Pedro, Túlio P. Navarro, Danilo Silva

Abstract: Pressure ulcers have high prevalence in ICU patients but are preventable if identified in initial stages. In practice, the Braden scale is used to classify high-risk patients. This paper investigates the use of machine learning in electronic health records data for this task, by using data available in MIMIC-III v1.4. Two main contributions are made: a new approach for evaluating models that consi… ▽ More Pressure ulcers have high prevalence in ICU patients but are preventable if identified in initial stages. In practice, the Braden scale is used to classify high-risk patients. This paper investigates the use of machine learning in electronic health records data for this task, by using data available in MIMIC-III v1.4. Two main contributions are made: a new approach for evaluating models that considers all predictions made during a stay, and a new training method for the machine learning models. The results show a superior performance in comparison to the state of the art; moreover, all models surpass the Braden scale in every operating point in the precision-recall curve. -- -- Lesões por pressão possuem alta prevalência em pacientes de UTI e são preveníveis ao serem identificadas em estágios iniciais. Na prática utiliza-se a escala de Braden para classificação de pacientes em risco. Este artigo investiga o uso de aprendizado de máquina em dados de registros eletrônicos para este fim, a partir da base de dados MIMIC-III v1.4. São feitas duas contribuições principais: uma nova abordagem para a avaliação dos modelos e da escala de Braden levando em conta todas as predições feitas ao longo das internações, e um novo método de treinamento para os modelos de aprendizado de máquina. Os resultados obtidos superam o estado da arte e verifica-se que os modelos superam significativamente a escala de Braden em todos os pontos de operação da curva de precisão por sensibilidade. △ Less

Submitted 23 December, 2021; originally announced December 2021.

Comments: 3 pages, 1 figure, in Portuguese, accepted at XVIII Congresso Brasileiro de Informática em Saúde (CBIS 2021)

FocusFace: Multi-task Contrastive Learning for Masked Face Recognition

Authors: Pedro C. Neto, Fadi Boutros, João Ribeiro Pinto, Naser Damer, Ana F. Sequeira, Jaime S. Cardoso

Abstract: SARS-CoV-2 has presented direct and indirect challenges to the scientific community. One of the most prominent indirect challenges advents from the mandatory use of face masks in a large number of countries. Face recognition methods struggle to perform identity verification with similar accuracy on masked and unmasked individuals. It has been shown that the performance of these methods drops consi… ▽ More SARS-CoV-2 has presented direct and indirect challenges to the scientific community. One of the most prominent indirect challenges advents from the mandatory use of face masks in a large number of countries. Face recognition methods struggle to perform identity verification with similar accuracy on masked and unmasked individuals. It has been shown that the performance of these methods drops considerably in the presence of face masks, especially if the reference image is unmasked. We propose FocusFace, a multi-task architecture that uses contrastive learning to be able to accurately perform masked face recognition. The proposed architecture is designed to be trained from scratch or to work on top of state-of-the-art face recognition methods without sacrificing the capabilities of a existing models in conventional face recognition tasks. We also explore different approaches to design the contrastive learning module. Results are presented in terms of masked-masked (M-M) and unmasked-masked (U-M) face verification performance. For both settings, the results are on par with published methods, but for M-M specifically, the proposed method was able to outperform all the solutions that it was compared to. We further show that when using our method on top of already existing methods the training computational costs decrease significantly while retaining similar performances. The implementation and the trained models are available at GitHub. △ Less

Submitted 1 November, 2021; v1 submitted 28 October, 2021; originally announced October 2021.

Comments: Accepted at the 16th IEEE International Conference on Automatic Face and Gesture Recognition, FG 2021

My Eyes Are Up Here: Promoting Focus on Uncovered Regions in Masked Face Recognition

Authors: Pedro C. Neto, Fadi Boutros, João Ribeiro Pinto, Mohsen Saffari, Naser Damer, Ana F. Sequeira, Jaime S. Cardoso

Abstract: The recent Covid-19 pandemic and the fact that wearing masks in public is now mandatory in several countries, created challenges in the use of face recognition systems (FRS). In this work, we address the challenge of masked face recognition (MFR) and focus on evaluating the verification performance in FRS when verifying masked vs unmasked faces compared to verifying only unmasked faces. We propose… ▽ More The recent Covid-19 pandemic and the fact that wearing masks in public is now mandatory in several countries, created challenges in the use of face recognition systems (FRS). In this work, we address the challenge of masked face recognition (MFR) and focus on evaluating the verification performance in FRS when verifying masked vs unmasked faces compared to verifying only unmasked faces. We propose a methodology that combines the traditional triplet loss and the mean squared error (MSE) intending to improve the robustness of an MFR system in the masked-unmasked comparison mode. The results obtained by our proposed method show improvements in a detailed step-wise ablation study. The conducted study showed significant performance gains induced by our proposed training paradigm and modified triplet loss on two evaluation databases. △ Less

Submitted 18 August, 2021; v1 submitted 2 August, 2021; originally announced August 2021.

Comments: Accepted at 20th International Conference of the Biometrics Special Interest Group (BIOSIG 2021)

MFR 2021: Masked Face Recognition Competition

Authors: Fadi Boutros, Naser Damer, Jan Niklas Kolf, Kiran Raja, Florian Kirchbuchner, Raghavendra Ramachandra, Arjan Kuijper, Pengcheng Fang, Chao Zhang, Fei Wang, David Montero, Naiara Aginako, Basilio Sierra, Marcos Nieto, Mustafa Ekrem Erakin, Ugur Demir, Hazim Kemal, Ekenel, Asaki Kataoka, Kohei Ichikawa, Shizuma Kubo, Jie Zhang, Mingjie He, Dan Han, Shiguang Shan , et al. (10 additional authors not shown)

Abstract: This paper presents a summary of the Masked Face Recognition Competitions (MFR) held within the 2021 International Joint Conference on Biometrics (IJCB 2021). The competition attracted a total of 10 participating teams with valid submissions. The affiliations of these teams are diverse and associated with academia and industry in nine different countries. These teams successfully submitted 18 vali… ▽ More This paper presents a summary of the Masked Face Recognition Competitions (MFR) held within the 2021 International Joint Conference on Biometrics (IJCB 2021). The competition attracted a total of 10 participating teams with valid submissions. The affiliations of these teams are diverse and associated with academia and industry in nine different countries. These teams successfully submitted 18 valid solutions. The competition is designed to motivate solutions aiming at enhancing the face recognition accuracy of masked faces. Moreover, the competition considered the deployability of the proposed solutions by taking the compactness of the face recognition models into account. A private dataset representing a collaborative, multi-session, real masked, capture scenario is used to evaluate the submitted solutions. In comparison to one of the top-performing academic face recognition solutions, 10 out of the 18 submitted solutions did score higher masked face verification accuracy. △ Less

Submitted 29 June, 2021; originally announced June 2021.

Comments: Accepted at International Join Conference on Biometrics (IJCB 2021)

Noncatastrophic convolutional codes over a finite ring

Authors: Diego Napp, Raquel Pinto, Conceição Rocha

Abstract: Noncatastrophic encoders are an important class of polynomial generator matrices of convolutional codes. When these polynomials have coefficients in a finite field, these encoders have been characterized are being polynomial left prime matrices. In this paper we study the notion of noncatastrophicity in the context of convolutional codes when the polynomial matrices have entries in a finite ring.… ▽ More Noncatastrophic encoders are an important class of polynomial generator matrices of convolutional codes. When these polynomials have coefficients in a finite field, these encoders have been characterized are being polynomial left prime matrices. In this paper we study the notion of noncatastrophicity in the context of convolutional codes when the polynomial matrices have entries in a finite ring. In particular, we need to introduce two different notion of primeness in order to fully characterize noncatastrophic encoders over the finite ring Z_{p^r}. The second part of the paper is devoted to investigate the notion of free and column distance in this context when the convolutional code is a free finitely generated Z_{p^r}-module. We introduce the notion of b-degree and provide new bounds on the free distances and column distance. We show that this class of convolutional codes is optimal with respect to the column distance and to the free distance if and only if its projection on Z_p is. △ Less

Submitted 14 April, 2021; originally announced April 2021.

Comments: 17 pages

VC Dimension and Distribution-Free Sample-Based Testing

Authors: Eric Blais, Renato Ferreira Pinto Jr., Nathaniel Harms

Abstract: We consider the problem of determining which classes of functions can be tested more efficiently than they can be learned, in the distribution-free sample-based model that corresponds to the standard PAC learning setting. Our main result shows that while VC dimension by itself does not always provide tight bounds on the number of samples required to test a class of functions in this model, it can… ▽ More We consider the problem of determining which classes of functions can be tested more efficiently than they can be learned, in the distribution-free sample-based model that corresponds to the standard PAC learning setting. Our main result shows that while VC dimension by itself does not always provide tight bounds on the number of samples required to test a class of functions in this model, it can be combined with a closely-related variant that we call "lower VC" (or LVC) dimension to obtain strong lower bounds on this sample complexity. We use this result to obtain strong and in many cases nearly optimal lower bounds on the sample complexity for testing unions of intervals, halfspaces, intersections of halfspaces, polynomial threshold functions, and decision trees. Conversely, we show that two natural classes of functions, juntas and monotone functions, can be tested with a number of samples that is polynomially smaller than the number of samples required for PAC learning. Finally, we also use the connection between VC dimension and property testing to establish new lower bounds for testing radius clusterability and testing feasibility of linear constraint systems. △ Less

Submitted 7 December, 2020; originally announced December 2020.

Comments: 44 pages

Fast Reinforcement Learning with Incremental Gaussian Mixture Models

Authors: Rafael Pinto

Abstract: This work presents a novel algorithm that integrates a data-efficient function approximator with reinforcement learning in continuous state spaces. An online and incremental algorithm capable of learning from a single pass through data, called Incremental Gaussian Mixture Network (IGMN), was employed as a sample-efficient function approximator for the joint state and Q-values space, all in a singl… ▽ More This work presents a novel algorithm that integrates a data-efficient function approximator with reinforcement learning in continuous state spaces. An online and incremental algorithm capable of learning from a single pass through data, called Incremental Gaussian Mixture Network (IGMN), was employed as a sample-efficient function approximator for the joint state and Q-values space, all in a single model, resulting in a concise and data-efficient algorithm, i.e., a reinforcement learning algorithm that learns from very few interactions with the environment. Results are analyzed to explain the properties of the obtained algorithm, and it is observed that the use of the IGMN function approximator brings some important advantages to reinforcement learning in relation to conventional neural networks trained by gradient descent methods. △ Less

Submitted 1 November, 2020; originally announced November 2020.

Comments: 17 pages, 8 figures

Model-Free Episodic Control with State Aggregation

Authors: Rafael Pinto

Abstract: Episodic control provides a highly sample-efficient method for reinforcement learning while enforcing high memory and computational requirements. This work proposes a simple heuristic for reducing these requirements, and an application to Model-Free Episodic Control (MFEC) is presented. Experiments on Atari games show that this heuristic successfully reduces MFEC computational demands while produc… ▽ More Episodic control provides a highly sample-efficient method for reinforcement learning while enforcing high memory and computational requirements. This work proposes a simple heuristic for reducing these requirements, and an application to Model-Free Episodic Control (MFEC) is presented. Experiments on Atari games show that this heuristic successfully reduces MFEC computational demands while producing no significant loss of performance when conservative choices of hyperparameters are used. Consequently, episodic control becomes a more feasible option when dealing with reinforcement learning tasks. △ Less

Submitted 21 August, 2020; originally announced August 2020.

Comments: 8 pages, 21 figures

List decoding of Convolutional Codes over integer residue rings

Authors: Julia Lieb, Diego Napp, Raquel Pinto

Abstract: A convolutional code $\C$ over $\ZZ[D]$ is a $\ZZ[D]$-submodule of $\ZZN[D]$ where $\ZZ[D]$ stands for the ring of polynomials with coefficients in $\ZZ$. In this paper, we study the list decoding problem of these codes when the transmission is performed over an erasure channel, that is, we study how much information one can recover from a codeword $w\in \C$ when some of its coefficients have been… ▽ More A convolutional code $\C$ over $\ZZ[D]$ is a $\ZZ[D]$-submodule of $\ZZN[D]$ where $\ZZ[D]$ stands for the ring of polynomials with coefficients in $\ZZ$. In this paper, we study the list decoding problem of these codes when the transmission is performed over an erasure channel, that is, we study how much information one can recover from a codeword $w\in \C$ when some of its coefficients have been erased. We do that using the $p$-adic expansion of $w$ and particular representations of the parity-check polynomial matrix of the code. From these matrix polynomial representations we recursively select certain equations that $w$ must satisfy and have only coefficients in the field $p^{r-1}\ZZ$. We exploit the natural block Toeplitz structure of the sliding parity-check matrix to derive a step by step methodology to obtain a list of possible codewords for a given corrupted codeword $w$, that is, a list with the closest codewords to $w$. △ Less

Submitted 8 September, 2020; v1 submitted 19 June, 2020; originally announced June 2020.

A decoding algorithm for 2D convolutional codes over the erasure channel

Authors: Julia Lieb, Raquel Pinto

Abstract: Two-dimensional (2D) convolutional codes are a generalization of (1D) convolutional codes, which are very appropriate for transmission over an erasure channel. In this paper, we present a decoding algorithm for 2D convolutional codes over this kind of channel by reducing the decoding process to several decoding steps with 1D convolutional codes. Moreover, we provide constructions of 2D convolution… ▽ More Two-dimensional (2D) convolutional codes are a generalization of (1D) convolutional codes, which are very appropriate for transmission over an erasure channel. In this paper, we present a decoding algorithm for 2D convolutional codes over this kind of channel by reducing the decoding process to several decoding steps with 1D convolutional codes. Moreover, we provide constructions of 2D convolutional codes that are specially taylored to our decoding algorithm. △ Less

Submitted 18 June, 2020; originally announced June 2020.

Delightful Companions: Supporting Well-Being Through Design Delight

Authors: Omar Sosa-Tzec, Gowri Balasubramaniam, Sylvia Sinsabaugh, Evan Sobetski, Rogerio Pinto, Shervin Assari

Abstract: This paper presents three design products referred to as delightful companions that are intended to help people engage in well-being practices. It also introduces the approach utilized to guide the design decisions during their creation. Design delight is the name of this approach, which comprises six experiential qualities that are regarded as antecedents of delight. The objective of this paper i… ▽ More This paper presents three design products referred to as delightful companions that are intended to help people engage in well-being practices. It also introduces the approach utilized to guide the design decisions during their creation. Design delight is the name of this approach, which comprises six experiential qualities that are regarded as antecedents of delight. The objective of this paper is to introduce the approach and the companions and state the two paths that have defined the future steps of this research. △ Less

Submitted 24 April, 2020; originally announced May 2020.

Comments: 5 pages, 9 Figures

ACM Class: H.5.m

Convolutional Codes

Authors: Julia Lieb, Raquel Pinto, Joachim Rosenthal

Abstract: The article provides a survey on convolutional codes stressing the connections to module theory and systems theory. Constructions of codes with maximal possible distance and distance profile are provided. The article will appear as book chapter in "A Concise Encyclopedia of Coding Theory" to be published by CRC Press. The article provides a survey on convolutional codes stressing the connections to module theory and systems theory. Constructions of codes with maximal possible distance and distance profile are provided. The article will appear as book chapter in "A Concise Encyclopedia of Coding Theory" to be published by CRC Press. △ Less

Submitted 22 January, 2020; originally announced January 2020.

Text-based inference of moral sentiment change

Authors: Jing Yi Xie, Renato Ferreira Pinto Jr., Graeme Hirst, Yang Xu

Abstract: We present a text-based framework for investigating moral sentiment change of the public via longitudinal corpora. Our framework is based on the premise that language use can inform people's moral perception toward right or wrong, and we build our methodology by exploring moral biases learned from diachronic word embeddings. We demonstrate how a parameter-free model supports inference of historica… ▽ More We present a text-based framework for investigating moral sentiment change of the public via longitudinal corpora. Our framework is based on the premise that language use can inform people's moral perception toward right or wrong, and we build our methodology by exploring moral biases learned from diachronic word embeddings. We demonstrate how a parameter-free model supports inference of historical shifts in moral sentiment toward concepts such as slavery and democracy over centuries at three incremental levels: moral relevance, moral polarity, and fine-grained moral dimensions. We apply this methodology to visualizing moral time courses of individual concepts and analyzing the relations between psycholinguistic variables and rates of moral sentiment change at scale. Our work offers opportunities for applying natural language processing toward characterizing moral sentiment change in society. △ Less

Submitted 20 January, 2020; originally announced January 2020.

Comments: In Proceedings of EMNLP 2019

Robust low-delay Streaming PIR using convolutional codes

Authors: Julia Lieb, Diego Napp, Raquel Pinto

Abstract: In this paper we investigate the design of a low-delay robust streaming PIR scheme on coded data that is resilient to unresponsive or slow servers and can privately retrieve streaming data in a sequential fashion subject to a fixed decoding delay. We present a scheme based on convolutional codes and the star product and assume no collusion between servers. In particular we propose the use of convo… ▽ More In this paper we investigate the design of a low-delay robust streaming PIR scheme on coded data that is resilient to unresponsive or slow servers and can privately retrieve streaming data in a sequential fashion subject to a fixed decoding delay. We present a scheme based on convolutional codes and the star product and assume no collusion between servers. In particular we propose the use of convolutional codes that have the maximum distance increase, called Maximum Distance Profile (MDP). We show that the proposed scheme can deal with many different erasure patterns. △ Less

Submitted 5 November, 2019; v1 submitted 4 November, 2019; originally announced November 2019.

Comments: 12 pages

Constructions of MDS convolutional codes using superregular matrices

Authors: Julia Lieb, Raquel Pinto

Abstract: Maximum distance separable convolutional codes are the codes that present best performance in error correction among all convolutional codes with certain rate and degree. In this paper, we show that taking the constant matrix coefficients of a polynomial matrix as submatrices of a superregular matrix, we obtain a column reduced generator matrix of an MDS convolutional code with a certain rate and… ▽ More Maximum distance separable convolutional codes are the codes that present best performance in error correction among all convolutional codes with certain rate and degree. In this paper, we show that taking the constant matrix coefficients of a polynomial matrix as submatrices of a superregular matrix, we obtain a column reduced generator matrix of an MDS convolutional code with a certain rate and a certain degree. We then present two novel constructions that fulfill these conditions by considering two types of superregular matrices. △ Less

Submitted 29 May, 2019; v1 submitted 26 March, 2019; originally announced March 2019.

MSC Class: 94B10

Ranking News-Quality Multimedia

Authors: Gonçalo Marcelino, Ricardo Pinto, João Magalhães

Abstract: News editors need to find the photos that best illustrate a news piece and fulfill news-media quality standards, while being pressed to also find the most recent photos of live events. Recently, it became common to use social-media content in the context of news media for its unique value in terms of immediacy and quality. Consequently, the amount of images to be considered and filtered through is… ▽ More News editors need to find the photos that best illustrate a news piece and fulfill news-media quality standards, while being pressed to also find the most recent photos of live events. Recently, it became common to use social-media content in the context of news media for its unique value in terms of immediacy and quality. Consequently, the amount of images to be considered and filtered through is now too much to be handled by a person. To aid the news editor in this process, we propose a framework designed to deliver high-quality, news-press type photos to the user. The framework, composed of two parts, is based on a ranking algorithm tuned to rank professional media highly and a visual SPAM detection module designed to filter-out low-quality media. The core ranking algorithm is leveraged by aesthetic, social and deep-learning semantic features. Evaluation showed that the proposed framework is effective at finding high-quality photos (true-positive rate) achieving a retrieval MAP of 64.5% and a classification precision of 70%. △ Less

Submitted 9 October, 2018; originally announced October 2018.

Comments: To appear in ICMR'18

ACM Class: H.3.3

Multi-Period Flexibility Forecast for Low Voltage Prosumers

Authors: Rui Pinto, Ricardo Bessa, Manuel Matos

Abstract: Near-future electric distribution grids operation will have to rely on demand-side flexibility, both by implementation of demand response strategies and by taking advantage of the intelligent management of increasingly common small-scale energy storage. The Home energy management system (HEMS), installed at low voltage residential clients, will play a crucial role on the flexibility provision to b… ▽ More Near-future electric distribution grids operation will have to rely on demand-side flexibility, both by implementation of demand response strategies and by taking advantage of the intelligent management of increasingly common small-scale energy storage. The Home energy management system (HEMS), installed at low voltage residential clients, will play a crucial role on the flexibility provision to both system operators and market players like aggregators. Modeling and forecasting multi-period flexibility from residential prosumers, such as battery storage and electric water heater, while complying with internal constraints (comfort levels, data privacy) and uncertainty is a complex task. This papers describes a computational method that is capable of efficiently learn and define the feasibility flexibility space from controllable resources connected to a HEMS. An Evolutionary Particle Swarm Optimization (EPSO) algorithm is adopted and reshaped to derive a set of feasible temporal trajectories for the residential net-load, considering storage, flexible appliances, and predefined costumer preferences, as well as load and photovoltaic (PV) forecast uncertainty. A support vector data description (SVDD) algorithm is used to build models capable of classifying feasible and non-feasible HEMS operating trajectories upon request from an optimization/control algorithm operated by a DSO or market player. △ Less

Submitted 8 November, 2017; v1 submitted 26 March, 2017; originally announced March 2017.

Scalable and Incremental Learning of Gaussian Mixture Models

Authors: Rafael Pinto, Paulo Engel

Abstract: This work presents a fast and scalable algorithm for incremental learning of Gaussian mixture models. By performing rank-one updates on its precision matrices and determinants, its asymptotic time complexity is of \BigO{NKD^2} for $N$ data points, $K$ Gaussian components and $D$ dimensions. The resulting algorithm can be applied to high dimensional tasks, and this is confirmed by applying it to th… ▽ More This work presents a fast and scalable algorithm for incremental learning of Gaussian mixture models. By performing rank-one updates on its precision matrices and determinants, its asymptotic time complexity is of \BigO{NKD^2} for $N$ data points, $K$ Gaussian components and $D$ dimensions. The resulting algorithm can be applied to high dimensional tasks, and this is confirmed by applying it to the classification datasets MNIST and CIFAR-10. Additionally, in order to show the algorithm's applicability to function approximation and control tasks, it is applied to three reinforcement learning tasks and its data-efficiency is evaluated. △ Less

Submitted 14 January, 2017; originally announced January 2017.

Comments: 13 pages, 1 figure, submitted for peer-review. arXiv admin note: substantial text overlap with arXiv:1506.04422

ACM Class: I.2.6

The dual of convolutional codes over $\mathbb{Z}_{p^r}$

Authors: Mohammed El Oued, Diego Napp, Raquel Pinto, Marisa Toste

Abstract: An important class of codes widely used in applications is the class of convolutional codes. Most of the literature of convolutional codes is devoted to con- volutional codes over finite fields. The extension of the concept of convolutional codes from finite fields to finite rings have attracted much attention in recent years due to fact that they are the most appropriate codes for phase modulatio… ▽ More An important class of codes widely used in applications is the class of convolutional codes. Most of the literature of convolutional codes is devoted to con- volutional codes over finite fields. The extension of the concept of convolutional codes from finite fields to finite rings have attracted much attention in recent years due to fact that they are the most appropriate codes for phase modulation. However convolutional codes over finite rings are more involved and not fully understood. Many results and features that are well-known for convolutional codes over finite fields have not been fully investigated in the context of finite rings. In this paper we focus in one of these unexplored areas, namely, we investigate the dual codes of convolutional codes over finite rings. In particular we study the p-dimension of the dual code of a convolutional code over a finite ring. This contribution can be considered a generalization and an extension, to the rings case, of the work done by Forney and McEliece on the dimension of the dual code of a convolutional code over a finite field. △ Less

Submitted 20 January, 2016; originally announced January 2016.

Comments: submitted

On MDS convolutional Codes over $\mathbb Z_{p^r}$

Authors: Diego Napp, Raquel Pinto, Marisa Toste

Abstract: Maximum Distance Separable (MDS) convolutional codes are cha- racterized through the property that the free distance meets the generalized Singleton bound. The existence of free MDS convolutional codes over Z p r was recently discovered in [26] via the Hensel lift of a cyclic code. In this paper we further investigate this important class of convolutional codes over Z p r from a new perspective. W… ▽ More Maximum Distance Separable (MDS) convolutional codes are cha- racterized through the property that the free distance meets the generalized Singleton bound. The existence of free MDS convolutional codes over Z p r was recently discovered in [26] via the Hensel lift of a cyclic code. In this paper we further investigate this important class of convolutional codes over Z p r from a new perspective. We introduce the notions of p-standard form and r- optimal parameters to derive a novel upper bound of Singleton type on the free distance. Moreover, we present a constructive method for building general (non necessarily free) MDS convolutional codes over Z p r for any given set of parameters. △ Less

Submitted 18 January, 2016; originally announced January 2016.

Superregular matrices and applications to convolutional codes

Authors: P. J. Almeida, D. Napp, R. Pinto

Abstract: The main results of this paper are twofold: the first one is a matrix theoretical result. We say that a matriz is superregular if all of its minors that are not trivially zero are nonzero. Given a a times b, a larger than or equal to b, superregular matrix over a field, we show that if all of its rows are nonzero then any linear combination of its columns, with nonzero coefficients, has at least a… ▽ More The main results of this paper are twofold: the first one is a matrix theoretical result. We say that a matriz is superregular if all of its minors that are not trivially zero are nonzero. Given a a times b, a larger than or equal to b, superregular matrix over a field, we show that if all of its rows are nonzero then any linear combination of its columns, with nonzero coefficients, has at least a-b+1 nonzero entries. Secondly, we make use of this result to construct convolutional codes that attain the maximum possible distance for some fixed parameters of the code, namely, the rate and the Forney indices. These results answer some open questions on distances and constructions of convolutional codes posted in the literature. △ Less

Submitted 12 January, 2016; originally announced January 2016.

MSC Class: 94B10; 15B33

A Fast Incremental Gaussian Mixture Model

Authors: Rafael Pinto, Paulo Engel

Abstract: This work builds upon previous efforts in online incremental learning, namely the Incremental Gaussian Mixture Network (IGMN). The IGMN is capable of learning from data streams in a single-pass by improving its model after analyzing each data point and discarding it thereafter. Nevertheless, it suffers from the scalability point-of-view, due to its asymptotic time complexity of… ▽ More This work builds upon previous efforts in online incremental learning, namely the Incremental Gaussian Mixture Network (IGMN). The IGMN is capable of learning from data streams in a single-pass by improving its model after analyzing each data point and discarding it thereafter. Nevertheless, it suffers from the scalability point-of-view, due to its asymptotic time complexity of $\operatorname{O}\bigl(NKD^3\bigr)$ for $N$ data points, $K$ Gaussian components and $D$ dimensions, rendering it inadequate for high-dimensional data. In this paper, we manage to reduce this complexity to $\operatorname{O}\bigl(NKD^2\bigr)$ by deriving formulas for working directly with precision matrices instead of covariance matrices. The final result is a much faster and scalable algorithm which can be applied to high dimensional tasks. This is confirmed by applying the modified algorithm to high-dimensional classification datasets. △ Less

Submitted 18 June, 2015; v1 submitted 14 June, 2015; originally announced June 2015.

Comments: 10 pages, no figures, draft submission to Plos One

ACM Class: I.2.6

A new class of superregular matrices and MDP convolutional codes

Authors: P. Almeida, D. Napp, R. Pinto

Abstract: This paper deals with the problem of constructing superregular matrices that lead to MDP convolutional codes. These matrices are a type of lower block triangular Toeplitz matrices with the property that all the square submatrices that can possibly be nonsingular due to the lower block triangular structure are nonsingular. We present a new class of matrices that are superregular over a suficiently… ▽ More This paper deals with the problem of constructing superregular matrices that lead to MDP convolutional codes. These matrices are a type of lower block triangular Toeplitz matrices with the property that all the square submatrices that can possibly be nonsingular due to the lower block triangular structure are nonsingular. We present a new class of matrices that are superregular over a suficiently large finite field F. Such construction works for any given choice of characteristic of the field F and code parameters (n; k; d) such that (n-k)|d. Finally, we discuss the size of F needed so that the proposed matrices are superregular. △ Less

Submitted 15 March, 2013; originally announced March 2013.

An iterative algorithm for parametrization of shortest length shift registers over finite rings

Authors: M. Kuijper, R. Pinto

Abstract: The construction of shortest feedback shift registers for a finite sequence S_1,...,S_N is considered over the finite ring Z_{p^r}. A novel algorithm is presented that yields a parametrization of all shortest feedback shift registers for the sequence of numbers S_1,...,S_N, thus solving an open problem in the literature. The algorithm iteratively processes each number, starting with S_1, and const… ▽ More The construction of shortest feedback shift registers for a finite sequence S_1,...,S_N is considered over the finite ring Z_{p^r}. A novel algorithm is presented that yields a parametrization of all shortest feedback shift registers for the sequence of numbers S_1,...,S_N, thus solving an open problem in the literature. The algorithm iteratively processes each number, starting with S_1, and constructs at each step a particular type of minimal Gröbner basis. The construction involves a simple update rule at each step which leads to computational efficiency. It is shown that the algorithm simultaneously computes a similar parametrization for the reciprocal sequence S_N,...,S_1. △ Less

Submitted 27 January, 2012; originally announced January 2012.

Comments: Submitted

MSC Class: 94A55; 11T71

Journal ref: Designs, Codes and Cryptography , pp. 1-23, 2016

On minimality of convolutional ring encoders

Authors: Margreta Kuijper, Raquel Pinto

Abstract: Convolutional codes are considered with code sequences modelled as semi-infinite Laurent series. It is wellknown that a convolutional code C over a finite group G has a minimal trellis representation that can be derived from code sequences. It is also wellknown that, for the case that G is a finite field, any polynomial encoder of C can be algebraically manipulated to yield a minimal polynomial… ▽ More Convolutional codes are considered with code sequences modelled as semi-infinite Laurent series. It is wellknown that a convolutional code C over a finite group G has a minimal trellis representation that can be derived from code sequences. It is also wellknown that, for the case that G is a finite field, any polynomial encoder of C can be algebraically manipulated to yield a minimal polynomial encoder whose controller canonical realization is a minimal trellis. In this paper we seek to extend this result to the finite ring case G = Z_{p^r} by introducing a socalled "p-encoder". We show how to manipulate a polynomial encoding of a noncatastrophic convolutional code over Z_{p^r} to produce a particular type of p-encoder ("minimal p-encoder") whose controller canonical realization is a minimal trellis with nonlinear features. The minimum number of trellis states is then expressed as p^gamma, where gamma is the sum of the row degrees of the minimal p-encoder. In particular, we show that any convolutional code over Z_{p^r} admits a delay-free p-encoder which implies the novel result that delay-freeness is not a property of the code but of the encoder, just as in the field case. We conjecture that a similar result holds with respect to catastrophicity, i.e., any catastrophic convolutional code over Z_{p^r} admits a noncatastrophic p-encoder. △ Less

Submitted 14 April, 2009; v1 submitted 24 January, 2008; originally announced January 2008.

Comments: 13 pages in v1, submitted; 8 pages in revision v2

Journal ref: IEEE Trans. Information Theory, Vol. 55, No. 11, pp. 4890-4897, November 2009

Line and Word Matching in Old Documents

Authors: A. Marcolino, Vitorino Ramos, Mario Ramalho, J. R. Caldas Pinto

Abstract: This paper is concerned with the problem of establishing an index based on word matching. It is assumed that the book was digitised as better as possible and some pre-processing techniques were already applied as line orientation correction and some noise removal. However two main factor are responsible for being not possible to apply ordinary optical character recognition techniques (OCR): the… ▽ More This paper is concerned with the problem of establishing an index based on word matching. It is assumed that the book was digitised as better as possible and some pre-processing techniques were already applied as line orientation correction and some noise removal. However two main factor are responsible for being not possible to apply ordinary optical character recognition techniques (OCR): the presence of antique fonts and the degraded state of many characters due to unrecoverable original time degradation. In this paper we make a short introduction to word segmentation that involves finding the lines that characterise a word. After we discuss different approaches for word matching and how they can be combined to obtain an ordered list for candidate words for the matching. This discussion will be illustrated by examples. △ Less

Submitted 17 December, 2004; originally announced December 2004.

Comments: 12 pages, 7 figures, Author at http://alfa.ist.utl.pt/~cvrm/staff/vramos/ref_32.html

ACM Class: I.2; I.5

Journal ref: SIARP 2000 - 5th IberoAmerican Symp. on Pattern Rec., F. Muge, Moises P. and R. Caldas Pinto (Eds.), ISBN 972-97711-1-1, pp. 123-135, Lisbon, Portugal, 11-13 Sep. 2000