Marcin Detyniecki

Bio: Marcin Detyniecki is an academic researcher from University of Paris. The author has contributed to research in topics: Computer science & Image retrieval. The author has an hindex of 9, co-authored 57 publications receiving 285 citations. Previous affiliations of Marcin Detyniecki include Polish Academy of Sciences & AXA.

Imperceptible Adversarial Attacks on Tabular Data

Abstract: Security of machine learning models is a concern as they may face adversarial attacks for unwarranted advantageous decisions. While research on the topic has mainly been focusing on the image domain, numerous industrial applications, in particular in finance, rely on standard tabular data. In this paper, we discuss the notion of adversarial examples in the tabular domain. We propose a formalization based on the imperceptibility of attacks in the tabular domain leading to an approach to generate imperceptible adversarial examples. Experiments show that we can generate imperceptible adversarial examples with a high fooling rate.

Ranking invariance between fuzzy similarity measures applied to image retrieval

Abstract: We first introduce the fuzzy similarity measures in the context of a CBIR system. This leads to the observation of an invariance in the ranking for different similarity measures. We then propose an explanation to this phenomenon, and a larger theory about order invariance for fuzzy similarity measures. We introduce a definition for equivalence classes based on order conservation between these measures. We then study the consequences of this theory on the evaluation of document retrieval by fuzzy similarity.

Fairness-Aware Neural R\'eyni Minimization for Continuous Features

Abstract: The past few years have seen a dramatic rise of academic and societal interest in fair machine learning. While plenty of fair algorithms have been proposed recently to tackle this challenge for discrete variables, only a few ideas exist for continuous ones. The objective in this paper is to ensure some independence level between the outputs of regression models and any given continuous sensitive variables. For this purpose, we use the Hirschfeld-Gebelein-Renyi (HGR) maximal correlation coefficient as a fairness metric. We propose two approaches to minimize the HGR coefficient. First, by reducing an upper bound of the HGR with a neural network estimation of the $\chi^{2}$ divergence. Second, by minimizing the HGR directly with an adversarial neural network architecture. The idea is to predict the output Y while minimizing the ability of an adversarial neural network to find the estimated transformations which are required to predict the HGR coefficient. We empirically assess and compare our approaches and demonstrate significant improvements on previously presented work in the field.

Detecting Adversarial Examples and Other Misclassifications in Neural Networks by Introspection.

Abstract: Despite having excellent performances for a wide variety of tasks, modern neural networks are unable to provide a reliable confidence value allowing to detect misclassifications. This limitation is at the heart of what is known as an adversarial example, where the network provides a wrong prediction associated with a strong confidence to a slightly modified image. Moreover, this overconfidence issue has also been observed for regular errors and out-of-distribution data. We tackle this problem by what we call introspection, i.e. using the information provided by the logits of an already pretrained neural network. We show that by training a simple 3-layers neural network on top of the logit activations, we are able to detect misclassifications at a competitive level.

Aggregation operators: properties, classes and construction methods

Abstract: Aggregation (fusion) of several input values into a single output value is an indispensable tool not only of mathematics or physics, but of majority of engineering, economical, social and other sciences. The problems of aggregation are very broad and heterogeneous, in general. Therefore we restrict ourselves in this contribution to the specific topic of the aggregation of finite number of real inputs only. Closely related topics of aggregating infinitely many real inputs [23,109,64,52,43,42,44,99], of aggregating inputs from some ordinal scales [41,50], of aggregating complex inputs (such as probability distributions [107,114], fuzzy sets [143]), etc., are treated, among others, in the quoted papers, and we will not deal with them. In this spirit, if the number of input values is fixed, say n, an aggregation operator is a real function of n variables. This is still a too general topic. Therefore we restrict our considerations regarding inputs as well as outputs to some fixed interval (scale) I = [a, b] ⊑ [-∞, ∞]. It is a matter of rescaling to fix I = [0,1].

The Fundamentalism of Liberal Rights: Decoding the Freedom of Expression Under the European Convention for the Protection of Human Rights and Fundamental Freedoms

Abstract: This paper attempts to deconstruct the free speech defense of the publications of cartoons offensive to many Muslims in Denmark and elsewhere in Europe in order to highlight the deep philosophical tensions between the characterizations of religion and race, between free speech and hate speech, and between the freedoms of expression and of religion. A scrutiny of the jurisprudence of the European Court of Human Rights (“ECtHR”) reveals the difficulties inherent in defining permissible limits on expression, particularly as it involves the identification and prioritization of interests that are worthy of protection under a state's law. The struggles over the characterization of certain interests as fundamental rights, in turn, raise questions over the ‘fundamental-ness' of rights and the valuation of foundational social and political values that the rhetoric of rights presumes as incontrovertible. This study seeks to advance the argument that fundamental rights, such as the freedom of expression, are legal constructs whose value is contingent on the ends they are employed to serve in a given socio-political environment. While the contingency of fundamental rights is palpable in debates over their definition and over what they include or exclude, it is most clearly visible in the clash of fundamental rights, in particular the freedoms of expression and religion.

Empirical Processes in M-Estimation

Abstract: (2001). Empirical Processes in M-Estimation. Journal of the American Statistical Association: Vol. 96, No. 454, pp. 779-780.

A survey on smartphone-based systems for opportunistic user context recognition

Abstract: The ever-growing computation and storage capability of mobile phones have given rise to mobile-centric context recognition systems, which are able to sense and analyze the context of the carrier so as to provide an appropriate level of service. As nonintrusive autonomous sensing and context recognition are desirable characteristics of a personal sensing system; efforts have been made to develop opportunistic sensing techniques on mobile phones. The resulting combination of these approaches has ushered in a new realm of applications, namely opportunistic user context recognition with mobile phones.This article surveys the existing research and approaches towards realization of such systems. In doing so, the typical architecture of a mobile-centric user context recognition system as a sequential process of sensing, preprocessing, and context recognition phases is introduced. The main techniques used for the realization of the respective processes during these phases are described, and their strengths and limitations are highlighted. In addition, lessons learned from previous approaches are presented as motivation for future research. Finally, several open challenges are discussed as possible ways to extend the capabilities of current systems and improve their real-world experience.