arXiv subjects

Jorge Silva

Publications and source records attributed to Jorge Silva.

16 recordsLinked to original sources

arXiv2025

SMOTE and Mirrors: Exposing Privacy Leakage from Synthetic Minority Oversampling

The Synthetic Minority Over-sampling Technique (SMOTE) is one of the most widely used methods for addressing class imbalance and generating synthetic data. Despite its popularity, little attention has been paid to its privacy implications; yet, it is used in the wild in many privacy-sensitive applications. In this work, we conduct the first systematic study of privacy leakage in SMOTE: we begin by showing that prevailing evaluation practices, i.e., naive distinguishing and distance-to-closest-record metrics, completely fail to detect any leakage and that membership inference attacks (MIAs) can be instantiated with high accuracy. Then, by exploiting SMOTE's geometric properties, we build two novel attacks with very limited assumptions: DistinSMOTE, which perfectly distinguishes real from synthetic records in augmented datasets, and ReconSMOTE, which reconstructs real minority records from synthetic datasets with perfect precision and recall approaching one under realistic imbalance ratios. We also provide theoretical guarantees for both attacks. Experiments on eight standard imbalanced datasets confirm the practicality and effectiveness of these attacks. Overall, our work reveals that SMOTE is inherently non-private and disproportionately exposes minority records, highlighting the need to reconsider its use in privacy-sensitive applications and as a baseline for assessing the privacy of modern generative models.

cs.CR↗

arXiv2025

dpmm: Differentially Private Marginal Models, a Library for Synthetic Tabular Data Generation

We propose dpmm, an open-source library for synthetic data generation with Differentially Private (DP) guarantees. It includes three popular marginal models -- PrivBayes, MST, and AIM -- that achieve superior utility and offer richer functionality compared to alternative implementations. Additionally, we adopt best practices to provide end-to-end DP guarantees and address well-known DP-related vulnerabilities. Our goal is to accommodate a wide audience with easy-to-install, highly customizable, and robust model implementations. Our codebase is available from https://github.com/sassoftware/dpmm.

cs.CR↗

arXiv2025

Robust Visual Servoing under Human Supervision for Assembly Tasks

We propose a framework enabling mobile manipulators to reliably complete pick-and-place tasks for assembling structures from construction blocks. The picking uses an eye-in-hand visual servoing controller for object tracking with Control Barrier Functions (CBFs) to ensure fiducial markers in the blocks remain visible. An additional robot with an eye-to-hand setup ensures precise placement, critical for structural stability. We integrate human-in-the-loop capabilities for flexibility and fault correction and analyze robustness to camera pose errors, proposing adapted barrier functions to handle them. Lastly, experiments validate the framework on 6-DoF mobile arms.

eess.SY↗

arXiv2024

Detectorless 3D terahertz imaging: achieving subwavelength resolution with reflectance confocal interferometric microscopy

Terahertz imaging holds great potential for non-destructive material inspection, but practical implementation has been limited by resolution constraints. In this study, we present a single-pixel THz imaging system based on a confocal microscope architecture, utilising a quantum cascade laser as both transmitter and phase-sensitive receiver. Our approach integrates laser feedback interferometry detection to achieve a two-fold improvement in lateral resolution and a two-order-of-magnitude enhancement in axial resolution over conventional imaging through precise interferometric phase measurements. This translates to a lateral resolution near $\lambda/2$ and a depth of focus better than $\lambda/5$, significantly outperforming traditional confocal systems. The system can produce a 0.5 Mpixel image in under two minutes, surpassing both raster-scanning single-pixel and multipixel focal-plane array-based imagers. Coherent operation enables simultaneous amplitude and phase image acquisition, and a custom visualisation method links amplitude to image saturation and phase to hue, enhancing material characterisation. A 3D tomographic analysis of a silicon chip reveals subwavelength features, demonstrating the system's potential for high-resolution THz imaging and material analysis. This work sets a new benchmark for THz imaging, overcoming key challenges and opening up transformative possibilities for non-destructive material inspection and characterisation.

physics.optics↗

arXiv2022

Machine Learning to Generate Adjustable Dose Distributions in Head-and-Neck Cancer Radiation Therapy

In this work, we propose a Machine Learning model that generates an adjustable 3D dose distribution for external beam radiation therapy for head-and-neck cancer treatments. In contrast to existing Machine Learning methods that provide a single model, we create pairs of models for each organ-at-risk, namely lower-extreme and upper-extreme models. These model pairs for an organ-at-risk propose doses that give lower and higher doses to that organ-at-risk, while also encapsulating the dose trade-off to other organs-at-risk. By weighting and combining the model pairs for all organs-at-risk, we are able to dynamically create adjustable dose distributions that can be used, in real-time, to move doses between organs-at-risk, thereby customizing the dose distribution to the needs of a particular patient. We leverage a key observation that the training data set inherently contains the clinical trade-offs. We show that the adjustable distributions are able to provide reasonable clinical dose latitude in the trade-off of doses between organs-at-risk.

physics.med-ph↗

arXiv2020

AttendLight: Universal Attention-Based Reinforcement Learning Model for Traffic Signal Control

We propose AttendLight, an end-to-end Reinforcement Learning (RL) algorithm for the problem of traffic signal control. Previous approaches for this problem have the shortcoming that they require training for each new intersection with a different structure or traffic flow distribution. AttendLight solves this issue by training a single, universal model for intersections with any number of roads, lanes, phases (possible signals), and traffic flow. To this end, we propose a deep RL model which incorporates two attention models. The first attention model is introduced to handle different numbers of roads-lanes; and the second attention model is intended for enabling decision-making with any number of phases in an intersection. As a result, our proposed model works for any intersection configuration, as long as a similar configuration is represented in the training set. Experiments were conducted with both synthetic and real-world standard benchmark data-sets. The results we show cover intersections with three or four approaching roads; one-directional/bi-directional roads with one, two, and three lanes; different number of phases; and different traffic flows. We consider two regimes: (i) single-environment training, single-deployment, and (ii) multi-environment training, multi-deployment. AttendLight outperforms both classical and other RL-based approaches on all cases in both regimes.

cs.LG↗

arXiv2018

The criticality of multiplicative processes

Keeping in view applications to numerical simulations of the evolution of a nuclear reactor core around criticality, we use a general mathematical framework for describing the evolutions of multiplicative processes (processes involving particle creation) both in particle generations and in time. This framework allows us to obtain, within a same formalism, two corresponding estimates of the multiplication factor which describes the growth of particle numbers at large times. We obtain the relative positions of both estimates with respect to each other and to criticality. These relations may show particularly useful when simulating in a realistic way the monitoring of nuclear cores in subcritical states, such as is the case for Accelerator Driven Systems (ADS). More generally, this study applies to various multiplicative processes which can be found in nature.

physics.gen-ph↗

arXiv2018

Multi-Task Learning with Incomplete Data for Healthcare

Multi-task learning is a type of transfer learning that trains multiple tasks simultaneously and leverages the shared information between related tasks to improve the generalization performance. However, missing features in the input matrix is a much more difficult problem which needs to be carefully addressed. Removing records with missing values can significantly reduce the sample size, which is impractical for datasets with large percentage of missing values. Popular imputation methods often distort the covariance structure of the data, which causes inaccurate inference. In this paper we propose using plug-in covariance matrix estimators to tackle the challenge of missing features. Specifically, we analyze the plug-in estimators under the framework of robust multi-task learning with LASSO and graph regularization, which captures the relatedness between tasks via graph regularization. We use the Alzheimer's disease progression dataset as an example to show how the proposed framework is effective for prediction and model estimation when missing data is present.

stat.ML↗

arXiv2018

RULLS: Randomized Union of Locally Linear Subspaces for Feature Engineering

Feature engineering plays an important role in the success of a machine learning model. Most of the effort in training a model goes into data preparation and choosing the right representation. In this paper, we propose a robust feature engineering method, Randomized Union of Locally Linear Subspaces (RULLS). We generate sparse, non-negative, and rotation invariant features in an unsupervised fashion. RULLS aggregates features from a random union of subspaces by describing each point using globally chosen landmarks. These landmarks serve as anchor points for choosing subspaces. Our method provides a way to select features that are relevant in the neighborhood around these chosen landmarks. Distances from each data point to $k$ closest landmarks are encoded in the feature matrix. The final feature representation is a union of features from all chosen subspaces. The effectiveness of our algorithm is shown on various real-world datasets for tasks such as clustering and classification of raw data and in the presence of noise. We compare our method with existing feature generation methods. Results show a high performance of our method on both classification and clustering tasks.

cs.LG↗

arXiv2017

Peak Criterion for Choosing Gaussian Kernel Bandwidth in Support Vector Data Description

Support Vector Data Description (SVDD) is a machine-learning technique used for single class classification and outlier detection. SVDD formulation with kernel function provides a flexible boundary around data. The value of kernel function parameters affects the nature of the data boundary. For example, it is observed that with a Gaussian kernel, as the value of kernel bandwidth is lowered, the data boundary changes from spherical to wiggly. The spherical data boundary leads to underfitting, and an extremely wiggly data boundary leads to overfitting. In this paper, we propose empirical criterion to obtain good values of the Gaussian kernel bandwidth parameter. This criterion provides a smooth boundary that captures the essential geometric features of the data.

cs.LG↗

arXiv2017

Online Robust Principal Component Analysis with Change Point Detection

Robust PCA methods are typically batch algorithms which requires loading all observations into memory before processing. This makes them inefficient to process big data. In this paper, we develop an efficient online robust principal component methods, namely online moving window robust principal component analysis (OMWRPCA). Unlike existing algorithms, OMWRPCA can successfully track not only slowly changing subspace but also abruptly changed subspace. By embedding hypothesis testing into the algorithm, OMWRPCA can detect change points of the underlying subspaces. Extensive simulation studies demonstrate the superior performance of OMWRPCA compared with other state-of-art approaches. We also apply the algorithm for real-time background subtraction of surveillance video.

cs.LG↗

arXiv2013

Analysis and interpretation of the Cramer-Rao lower-bound in astrometry: One dimensional case

In this paper we explore the maximum precision attainable in the location of a point source imaged by a pixel array detector in the presence of a background, as a function of the detector properties. For this we use a well-known result from parametric estimation theory, the so-called Cramer-Rao lower bound. We develop the expressions in the 1-dimensional case of a linear array detector in which the only unknown parameter is the source position. If the object is oversampled by the detector, analytical expressions can be obtained for the Cramer-Rao limit that can be readily used to estimate the limiting precision of an imaging system, and which are very useful for experimental (detector) design, observational planning, or performance estimation of data analysis software: In particular, we demonstrate that for background-dominated sources, the maximum astrometric precision goes as $B/F^2$, where $B$ is the background in one pixel, and $F$ is the total flux of the source, while when the background is negligible, this precision goes as $F^{-1}$. We also explore the dependency of the astrometric precision on: (1) the size of the source (as imaged by the detector), (2) the pixel detector size, and (3) the effect of source de-centering. Putting these results into context, the theoretical Cramer-Rao lower bound is compared to both ground- as well as spaced-based astrometric results, indicating that current techniques approach this limit very closely. Our results indicate that we have found in the Cramer-Rao lower variance bound a very powerful astrometric "benchmark" estimator concerning the maximum expected positional precision for a point source, given a prescription for the source, the background, the detector characteristics, and the detection process.

astro-ph.IM↗

arXiv2012

A recursive procedure for density estimation on the binary hypercube

This paper describes a recursive estimation procedure for multivariate binary densities (probability distributions of vectors of Bernoulli random variables) using orthogonal expansions. For $d$ covariates, there are $2^d$ basis coefficients to estimate, which renders conventional approaches computationally prohibitive when $d$ is large. However, for a wide class of densities that satisfy a certain sparsity condition, our estimator runs in probabilistic polynomial time and adapts to the unknown sparsity of the underlying density in two key ways: (1) it attains near-minimax mean-squared error for moderate sample sizes, and (2) the computational complexity is lower for sparser densities. Our method also allows for flexible control of the trade-off between mean-squared error and computational complexity.

math.ST↗

arXiv2012

Sequential anomaly detection in the presence of noise and limited feedback

This paper describes a methodology for detecting anomalies from sequentially observed and potentially noisy data. The proposed approach consists of two main elements: (1) {\em filtering}, or assigning a belief or likelihood to each successive measurement based upon our ability to predict it from previous noisy observations, and (2) {\em hedging}, or flagging potential anomalies by comparing the current belief against a time-varying and data-adaptive threshold. The threshold is adjusted based on the available feedback from an end user. Our algorithms, which combine universal prediction with recent work on online convex programming, do not require computing posterior distributions given all current observations and involve simple primal-dual parameter updates. At the heart of the proposed approach lie exponential-family models which can be used in a wide variety of contexts and applications, and which yield methods that achieve sublinear per-round regret against both static and slowly varying product distributions with marginals drawn from the same exponential family. Moreover, the regret against static distributions coincides with the minimax value of the corresponding online strongly convex game. We also prove bounds on the number of mistakes made during the hedging step relative to the best offline choice of the threshold with access to all estimated beliefs and feedback signals. We validate the theory on synthetic data drawn from a time-varying distribution over binary vectors of high dimensionality, as well as on the Enron email dataset.

cs.LG↗

arXiv2011

Blind Compressed Sensing Over a Structured Union of Subspaces

This paper addresses the problem of simultaneous signal recovery and dictionary learning based on compressive measurements. Multiple signals are analyzed jointly, with multiple sensing matrices, under the assumption that the unknown signals come from a union of a small number of disjoint subspaces. This problem is important, for instance, in image inpainting applications, in which the multiple signals are constituted by (incomplete) image patches taken from the overall image. This work extends standard dictionary learning and block-sparse dictionary optimization, by considering compressive measurements, e.g., incomplete data). Previous work on blind compressed sensing is also generalized by using multiple sensing matrices and relaxing some of the restrictions on the learned dictionary. Drawing on results developed in the context of matrix completion, it is proven that both the dictionary and signals can be recovered with high probability from compressed measurements. The solution is unique up to block permutations and invertible linear transformations of the dictionary atoms. The recovery is contingent on the number of measurements per signal and the number of signals being sufficiently large; bounds are derived for these quantities. In addition, this paper presents a computationally practical algorithm that performs dictionary learning and signal recovery, and establishes conditions for its convergence to a local optimum. Experimental results for image inpainting demonstrate the capabilities of the method.

cs.IT↗

arXiv2002

Preliminary Study of the Feasibility of a Non Crystalline Positron Emission Tomography Using a Suspension of Superheated Superconducting Grains (SSG) in High Density Dielectric Matrix (HDDM) as Detector

Suspensions of superheated superconducting grains are a detecting composite material. Each grain in the supension is a microcalorimeter with an energy threshold defined by its equatorial magnetic field for a given temperature. The higher the matrix density, the larger the gamma stopping power. For several years, cylindrical cells of such suspensions about 2 cm long and 0.4 mm in diameter can be read out in real time. As a result, using two independent cells, one can record a time coincidence between them. This could be potentially very useful for positron cameras where two diametrically opposite cells are simultaneously knocked by 511 keV gammas. This paper, based on the state of art in SSG in high density matrix, discusses such a feasibility.

physics.ins-det↗