Mário de Castro

Matheus Barreto, Mário de Castro, Thiago R. Ramos et al.

Modern machine learning models can be accurate on average yet still make mistakes that dominate deployment cost. We introduce Locus, a distribution-free wrapper that produces a per-input loss-scale reliability score for a fixed prediction function. Rather than quantifying uncertainty about the label, Locus models the realized loss of the prediction function using any engine that outputs a predictive distribution for the loss given an input. A simple split-calibration step turns this function into a distribution-free interpretable score that is comparable across inputs and can be read as an upper loss level. The score is useful on its own for ranking, and it can optionally be thresholded to obtain a transparent flagging rule with distribution-free control of large-loss events. Experiments across 13 regression benchmarks show that Locus yields effective risk ranking and reduces large-loss frequency compared to standard heuristics.

Matheus Vinícius Barreto de Farias, Mario de Castro

This study investigates the impact of adding noise to the training set classes in classification tasks using the BCOPS algorithm (Balanced and Conformal Optimized Prediction Sets), proposed by Guan & Tibshirani (2022). The BCOPS algorithm is an application of conformal prediction combined with a machine learning method to construct prediction sets such that the probability of the true class being included in the prediction set for a test observation meets a specified coverage guarantee. An observation is considered an outlier if its true class is not present in the training set. The study employs both synthetic and real datasets and conducts experiments to evaluate the prediction abstention rate for outlier observations and the model's robustness in this previously untested scenario. The results indicate that the addition of noise, even in small amounts, can have a significant effect on model performance.

João Batista Florindo, Mário de Castro, Odemir Martinez Bruno

This work proposes and study the concept of Functional Data Analysis transform, applying it to the performance improving of volumetric Bouligand-Minkowski fractal descriptors. The proposed transform consists essentially in changing the descriptors originally defined in the space of the calculus of fractal dimension into the space of coefficients used in the functional data representation of these descriptors. The transformed decriptors are used here in texture classification problems. The enhancement provided by the FDA transform is measured by comparing the transformed to the original descriptors in terms of the correctness rate in the classification of well known datasets.