Samuel Sousa

Mr. Samuel Sousa

orcid

Research Focus & Metrics

Academic Career Length

8Years (2019-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
21 May 2022 in Artificial Intelligence Review
Verified
Journal Article
Open access
PDF file
SSSamuel Sousa
Samuel Sousa
Roman KernRoman Kern
Roman Kern
Deep learning (DL) models for natural language processing (NLP) tasks often handle private data, demanding protection against breaches and disclosures. Data protection laws, such as the European Union’s General Data Protection Regulation (GDPR), thereby enforce the need for privacy. Although many privacy-preserving NLP methods have been proposed in recent years, no categories to organize them have been introduced yet, making it hard to follow the progress of the literature. To close this gap, this article systematically reviews over sixty DL methods for privacy-preserving NLP published between 2016 and 2020, covering theoretical foundations, privacy-enhancing technologies, and analysis of their suitability for real-world scenarios. First, we introduce a novel taxonomy for classifying the existing methods into three categories: data safeguarding methods, trusted methods, and verification methods. Second, we present an extensive summary of privacy threats, datasets for applications, and metrics for privacy evaluation. Third, throughout the review, we describe privacy issues in the NLP pipeline in a holistic view. Further, we discuss open challenges in privacy-preserving NLP regarding data traceability, computation overhead, dataset size, the prevalence of human biases in embeddings, and the privacy-utility tradeoff. Finally, this review presents future research directions to guide successive research and development of privacy-preserving NLP models.

6 February 2025 in Information
Verified
Journal Article
Open access
PDF file
SSSamuel Sousa
Samuel Sousa
MJMichael Jantscher
Michael Jantscher
MKMark Kröll
Mark Kröll
Roman KernRoman Kern
Roman Kern
Electronic health record (EHR) de-identification is crucial for publishing or sharing medical data without violating the patient’s privacy. Protected health information (PHI) is abundant in EHRs, and privacy regulations worldwide mandate de-identification before downstream tasks are performed. The ever-growing data generation in healthcare and the advent of generative artificial intelligence have increased the demand for de-identified EHRs and highlighted privacy issues with large language models (LLMs), especially data transmission to cloud-based LLMs. In this study, we benchmark ten LLMs for de-identifying EHRs in English and German. We then compare de-identification performance for in-context learning and full model fine-tuning and analyze the limitations of LLMs for this task. Our experimental evaluation shows that LLMs effectively de-identify EHRs in both languages. Moreover, in-context learning with a one-shot setting boosts de-identification performance without the costly full fine-tuning of the LLMs.

26 November 2024 in Neural Computing and Applications
Verified
Journal Article
Closed access
PDF file
SSSamuel Sousa
Samuel Sousa
ATAndreas Trügler
Andreas Trügler
Roman KernRoman Kern
Roman Kern

1 December 2019 in CLEI Electronic Journal
Verified
Journal Article
Open access
PDF file
SSSamuel Sousa
Samuel Sousa
RDRonaldo de Castro Del-Fiaco
Ronaldo de Castro Del-Fiaco
LBLilian Berton
Lilian Berton
Homicide is recognized as one of the most violent types of crime. In some countries, it is a hard problem to tackle because of its high occurrence and the lack of research on it. In Brazil, this problem is even harder, since this country is responsible for about 10% of the homicides in the world. Some Brazilian states suffer from the rise of homicide rates, like the state of Goi´as, in which its homicide rate increased from 24.5 per 100,000 in 2002 to 42.6 per 100,000 in 2014, becoming one of the five most violent states of Brazil, despite of having few population. This paper aims at applying clustering algorithms and feature selection models on criminal data concerning homicides and socio-economic variables in the state of Goi´as. We employed three clustering algorithms: K-means, Densitybased, and Hierarchical; as well as two feature selection models: Univariate Selection and Feature Importance. Our results indicate that homicide rates are more recurrent in large urban centers, although these cities have the best socio-economic indicators. Population and the educational level of the adult population were the variables which most influenced the results. K-means clustering brought the optimum outcomes, and Univariate Selection better selected attributes of the database.

20 October 2022 in arXiv
Verified
Preprint
Open access
PDF file
RWRoman Walch
Roman Walch
SSSamuel Sousa
Samuel Sousa
LHLukas Helminger
Lukas Helminger
SLStefanie Lindstaedt
Stefanie Lindstaedt
CRChristian Rechberger
Christian Rechberger
ATAndreas Trügler
Andreas Trügler
Big data has been a pervasive catchphrase in recent years, but dealing with data scarcity has become a crucial question for many real-world deep learning (DL) applications. A popular methodology to efficiently enable the training of DL models to perform tasks in scenarios with low availability of data is transfer learning (TL). TL allows to transfer knowledge from a general domain to a specific target one. However, such a knowledge transfer may put privacy at risk when it comes to sensitive or private data. With CryptoTL we introduce a solution to this problem, and show for the first time a cryptographic privacy-preserving TL approach based on homomorphic encryption that is efficient and feasible for real-world use cases. We achieve this by carefully designing the framework such that training is always done in plain while still profiting from the privacy gained by homomorphic encryption. To demonstrate the efficiency of our framework, we instantiate it with the popular CKKS HE scheme and apply CryptoTL to classification tasks with small datasets and show the applicability of our approach for sentiment analysis and spam detection. Additionally, we highlight how our approach can be combined with differential privacy to further increase the security guarantees. Our extensive benchmarks show that using CryptoTL leads to high accuracy while still having practical fine-tuning and classification runtimes despite using homomorphic encryption. Concretely, one forward-pass through the encrypted layers of our setup takes roughly 1s on a notebook CPU.

4 April 2022 in arXiv
Verified
Preprint
Open access
PDF file
SSSamuel Sousa
Samuel Sousa
CGChristian Guetl
Christian Guetl
RKRoman Kern
Roman Kern
Privacy is of worldwide concern regarding activities and processes that include sensitive data. For this reason, many countries and territories have been recently approving regulations controlling the extent to which organizations may exploit data provided by people. Artificial intelligence areas, such as machine learning and natural language processing, have already successfully employed privacy-preserving mechanisms in order to safeguard data privacy in a vast number of applications. Information retrieval (IR) is likewise prone to privacy threats, such as attacks and unintended disclosures of documents and search history, which may cripple the security of users and be penalized by data protection laws. This work aims at highlighting and discussing open challenges for privacy in the recent literature of IR, focusing on tasks featuring user-generated text data. Our contribution is threefold: firstly, we present an overview of privacy threats to IR tasks; secondly, we discuss applicable privacy-preserving mechanisms which may be employed in solutions to restrain privacy hazards; finally, we bring insights on the tradeoffs between privacy preservation and utility performance for IR tasks.