Hasil untuk "Diplomatics. Archives. Seals"

Shuhao Zhang, Yuli Chen, Jiale Han et al.

Watermarking provides a critical safeguard for large language model (LLM) services by facilitating the detection of LLM-generated text. Correspondingly, stealing watermark algorithms (SWAs) derive watermark information from watermarked texts generated by victim LLMs to craft highly targeted adversarial attacks, which compromise the reliability of watermarks. Existing SWAs rely on fixed strategies, overlooking the non-uniform distribution of stolen watermark information and the dynamic nature of real-world LLM generation processes. To address these limitations, we propose Adaptive Stealing (AS), a novel SWA featuring enhanced design flexibility through Position-Based Seal Construction and Adaptive Selection modules. AS operates by defining multiple attack perspectives derived from distinct activation states of contextually ordered tokens. During attack execution, AS dynamically selects the optimal perspective based on watermark compatibility, generation priority, and dynamic generation relevance. Our experiments demonstrate that AS significantly increases steal efficiency against target watermarks under identical experimental conditions. These findings highlight the need for more robust LLM watermarks to withstand potential attacks. We release our code to the community for future research\footnote{https://github.com/DrankXs/AdaptiveStealingWatermark}.

Francisco Vilanova, Luana do Nascimento Cabral

O presente artigo analisa a instrução pública primária no Piauí provincial, a partir da condição dos professores e da sua relação com a fiscalização escolar. Trata-se de uma pesquisa documental cujas fontes mobilizadas foram regulamentos, resoluções, relatórios da diretoria de instrução, além de matérias de jornais do período examinado. Os resultados revelam aspectos do processo de organização do ensino e as tensões entre professores e inspetores no contexto local. Palavras-chave: instrução pública; inspeção escolar; professor; Piauí provincial.

Andre Opris

This paper presents a first mathematical runtime analysis of PAES-25, an enhanced version of the original Pareto Archived Evolution Strategy (PAES) coming from the study of telecommunication problems over two decades ago to understand the dynamics of local search of MOEAs on many-objective fitness landscapes. We derive tight expected runtime bounds of PAES-25 with one-bit mutation on $m$-LOTZ until the entire Pareto front is found: $Θ(n^3)$ iterations if $m=2$, $Θ(n^3 \log^2(n))$ iterations if $m=4$ and $Θ(n(2n/m)^{m/2} \log(n/m))$ iterations if $m>4$ where $n$ is the problem size and $m$ the number of objectives. To the best of our knowledge, these are the first known tight runtime bounds for an MOEA outperforming the best known upper bound of $O(n^{m+1})$ for (G)SEMO on $m$-LOTZ when $m$ is at least $4$. We also show that archivers, such as the Adaptive Grid Archiver (AGA), Hypervolume Archiver (HVA) or Multi-Level Grid Archiver (MGA), help to distribute the set of solutions across the Pareto front of $m$-LOTZ efficiently. We also show that PAES-25 with standard bit mutation optimizes the bi-objective LOTZ benchmark in expected $O(n^4)$ iterations, and we discuss its limitations on other benchmarks such as OMM or COCZ.

Hector Giacomini

This note offers an overview of how Josiah Willard Gibbs's Elementary Principles in Statistical Mechanics, published simultaneously in London and New York in 1902, spread through European university libraries. Contrary to the received idea that the circulation of this text was slow, information gathered through direct contacts with numerous academic libraries, together with an examination of Yale University's archives, reveals an unexpectedly rapid material diffusion beginning on 15 March 1902. This early propagation can be explained by several channels: presentation copies sent by Yale University to leading universities, personal mailings by Gibbs himself to prominent scientists, and the distribution of copies by the American publisher to major scientific journals.

Paul Ahidjo

Depuis la période précoloniale, l’Extrême-Nord du Cameroun subit les affres climatiques. Les sécheresses y sévissent de façon récurrente, affectent les activités socio-économiques et provoquent la migration des populations vers des zones dites utiles. En s’intéressant aux impacts des crises écologiques des décennies 1970, 1980 et 1996, l’ambition de ce travail est de montrer comment les populations de l’Extrême-Nord du Cameroun ont perçu et ont réagi face aux sécheresses. Le déplacement des populations s’analyse et s’appréhende comme une forme d’adaptation aux risques climatiques. Dès lors, nous abordons tour à tour la perception des crises écologiques par les populations, les nouveaux comportements que provoquent les sécheresses chez les populations et enfin, la migration comme stratégie d’adaptation.

Paulina Krzywda

Lucimar Felisberto dos Santos

O artigo discute como as rotas que constituíram o sistema de circulação terrestre nos períodos colonial e imperial, necessárias à manutenção das atividades de abastecimento, comércio e trânsito de pessoas da região das Minas Gerais à cidade do Rio de Janeiro, contribuíram na execução do tráfico atlântico de africanos escravizados. Considerando o impacto da existência de portos de recepção clandestina na Baía da Guanabara, apresenta uma proposta de reconstituição das rotas negras mageenses. Palavras-chave: rotas negras; diáspora africana; comunidades tradicionais em Magé.

Attilio Bartoli Langeli

É correto que o cômpito de fazer a abertura um congresso tão denso e desafiador como este seja realizado da maneira mais serena, leve e fluida possível. Por isso, o propósito do conferencista é narrar, de forma muito breve, o que ocorreu no último século sob os dois céus da paleografia e da diplomática. A paleografia, ou seja, a história da escrita, mudou profundamente, embora tenha permanecido fiel ao seu princípio, que é o da análise formal dos produtos gráficos. O último ensinamento foi o de Armando Petrucci, falecido em 2018, que conseguiu ao mesmo tempo um fortíssimo renovamento dos estudos e uma grande valorização das metodologias próprias da paleografia. Já na diplomática, passou-se de um máximo de otimismo, o da escola positiva, a um máximo de pessimismo, ou seja, de negacionismo, o da nouvelle histoire; e talvez hoje seja o momento maduro para refletir sobre um dado objetivo: o documento. Ele não é nem um atalho nem uma renúncia em relação ao conhecimento da realidade do passado.

Eleanor Chodroff, Blaž Pažon, Annie Baker et al.

Research in speech technologies and comparative linguistics depends on access to diverse and accessible speech data. The UCLA Phonetics Lab Archive is one of the earliest multilingual speech corpora, with long-form audio recordings and phonetic transcriptions for 314 languages (Ladefoged et al., 2009). Recently, 95 of these languages were time-aligned with word-level phonetic transcriptions (Li et al., 2021). Here we present VoxAngeles, a corpus of audited phonetic transcriptions and phone-level alignments of the UCLA Phonetics Lab Archive, which uses the 95-language CMU re-release as our starting point. VoxAngeles also includes word- and phone-level segmentations from the original UCLA corpus, as well as phonetic measurements of word and phone durations, vowel formants, and vowel f0. This corpus enhances the usability of the original data, particularly for quantitative phonetic typology, as demonstrated through a case study of vowel intrinsic f0. We also discuss the utility of the VoxAngeles corpus for general research and pedagogy in crosslinguistic phonetics, as well as for low-resource and multilingual speech technologies. VoxAngeles is free to download and use under a CC-BY-NC 4.0 license.

Amit Das, Zheng Zhang, Najib Hasan et al.

Data annotation, the practice of assigning descriptive labels to raw data, is pivotal in optimizing the performance of machine learning models. However, it is a resource-intensive process susceptible to biases introduced by annotators. The emergence of sophisticated Large Language Models (LLMs) presents a unique opportunity to modernize and streamline this complex procedure. While existing research extensively evaluates the efficacy of LLMs, as annotators, this paper delves into the biases present in LLMs when annotating hate speech data. Our research contributes to understanding biases in four key categories: gender, race, religion, and disability with four LLMs: GPT-3.5, GPT-4o, Llama-3.1 and Gemma-2. Specifically targeting highly vulnerable groups within these categories, we analyze annotator biases. Furthermore, we conduct a comprehensive examination of potential factors contributing to these biases by scrutinizing the annotated data. We introduce our custom hate speech detection dataset, HateBiasNet, to conduct this research. Additionally, we perform the same experiments on the ETHOS (Mollas et al. 2022) dataset also for comparative analysis. This paper serves as a crucial resource, guiding researchers and practitioners in harnessing the potential of LLMs for data annotation, thereby fostering advancements in this critical field.

Iori Khuhro, Erin Gilmore, Jim Suderman et al.

As records are increasingly born digital – and thus, at least ostensibly, potentially much more accessible – archivists find themselves struggling to enable general access while providing appropriate privacy protections for the torrent of records being transferred to their care. In this article, the authors report the results of an integrative literature review study, examining the intersection of AI, archives, and privacy in terms of how archives are currently coping with these challenges and what role(s) AI might play in addressing privacy in archival records. The study revealed three major themes: 1) the challenges of – and possibilities beyond – defining “privacy” and “AI”; 2) the need for context-sensitive ways to manage privacy and access decisions; and 3) the lack of adequate “success measures” for ensuring the actual fitness for purpose of privacy AI solutions in the archival context.

Luiz Fernando Saraiva, Rita Almico, Thiago Campos Pessoa

O período entre 1831 e 1850 foi marcado pela entrada massiva de escravizados no Brasil, não obstante a sua proibição pelo governo imperial. Os impactos econômicos foram, até pouco tempo, pouco estudados pela historiografia. O artigo investiga a fortuna de José Bernardino de Sá, barão e visconde de Villa Nova do Minho, um dos mais ricos “capitalistas” da cidade do Rio de Janeiro em meados do século XIX, que teve a sua fortuna diretamente ligada à prática do comércio legal – e depois ilegal – de africanos. Palavras-chave: tráfico ilegal; escravidão; fortunas; investimentos.

Cyril Pernet, Claus Svarer, Ross Blair et al.

Accessing research data at any time is what FAIR (Findable Accessible Interoperable Reusable) data sharing aims to achieve at scale. Yet, we argue that it is not sustainable to keep accumulating and maintaining all datasets for rapid access, considering the monetary and ecological cost of maintaining repositories. Here, we address the issue of cold data storage: when to dispose of data for offline storage, how can this be done while maintaining FAIR principles and who should be responsible for cold archiving and long-term preservation.

Maxence Faldor, Félix Chalumeau, Manon Flageat et al.

Quality-Diversity algorithms, such as MAP-Elites, are a branch of Evolutionary Computation generating collections of diverse and high-performing solutions, that have been successfully applied to a variety of domains and particularly in evolutionary robotics. However, MAP-Elites performs a divergent search based on random mutations originating from Genetic Algorithms, and thus, is limited to evolving populations of low-dimensional solutions. PGA-MAP-Elites overcomes this limitation by integrating a gradient-based variation operator inspired by Deep Reinforcement Learning which enables the evolution of large neural networks. Although high-performing in many environments, PGA-MAP-Elites fails on several tasks where the convergent search of the gradient-based operator does not direct mutations towards archive-improving solutions. In this work, we present two contributions: (1) we enhance the Policy Gradient variation operator with a descriptor-conditioned critic that improves the archive across the entire descriptor space, (2) we exploit the actor-critic training to learn a descriptor-conditioned policy at no additional cost, distilling the knowledge of the archive into one single versatile policy that can execute the entire range of behaviors contained in the archive. Our algorithm, DCG-MAP-Elites improves the QD score over PGA-MAP-Elites by 82% on average, on a set of challenging locomotion tasks.

Xiangyu Wang, Jingsen Zhu, Qi Ye et al.

With the popularity of implicit neural representations, or neural radiance fields (NeRF), there is a pressing need for editing methods to interact with the implicit 3D models for tasks like post-processing reconstructed scenes and 3D content creation. While previous works have explored NeRF editing from various perspectives, they are restricted in editing flexibility, quality, and speed, failing to offer direct editing response and instant preview. The key challenge is to conceive a locally editable neural representation that can directly reflect the editing instructions and update instantly. To bridge the gap, we propose a new interactive editing method and system for implicit representations, called Seal-3D, which allows users to edit NeRF models in a pixel-level and free manner with a wide range of NeRF-like backbone and preview the editing effects instantly. To achieve the effects, the challenges are addressed by our proposed proxy function mapping the editing instructions to the original space of NeRF models in the teacher model and a two-stage training strategy for the student model with local pretraining and global finetuning. A NeRF editing system is built to showcase various editing types. Our system can achieve compelling editing effects with an interactive speed of about 1 second.

Fabian Huch

The Isabelle Archive of Formal Proofs has grown to a significant size in the past years. It makes up for an impressive body of research, which enables a number of statistical approaches to various aspects in theorem proving, and has not yet been utilized exhaustively. However, the growing size also poses some challenges to address: Material becomes increasingly harder to find, reusability and ease of understanding become more important. This thesis abstract summarizes my research plans on those topics and briefly touches on preliminary results, which indicate that the node in-degree of the dependency graph of the archive follows a scale-free distribution.

A. Nasidi Nadir

A Biographical Sketch of Reverend Jonathan C. Yep, 1934-2018 (Vestiges Biographical Notes Series)

Masayuki Tanaka, Hiroyuki Ikeda, Kazumi Murata et al.

We present the launch of the Hyper Suprime-Cam Legacy Archive (HSCLA), a public archive of processed, science-ready data from Hyper Suprime-Cam (HSC). HSC is an optical wide-field imager installed at the prime focus of the Subaru Telescope and has been in operation since 2014. While ~1/3 of the total observing time of HSC has been used for the Subaru Strategic Program (SSP), the remainder of the time is used for PI programs. We have processed the data from these PI programs and make the processed, high quality data available to the community through HSCLA. The current version of HSCLA includes data taken in the first year of science operation, 2014. We provide both individual and coadd images as well as photometric catalogs. The photometric catalog from the coadd is loaded to the database, which offers a fast access to the large catalog. There are other online tools such as image browser and image cutout tool and they will be useful for science analyses. The coadd images reach 24-27th magnitudes at $5σ$ for point sources and cover approximately 580 square degrees in at least one filter with 150 million objects in total. We perform extensive quality assurance tests and verify the photometric and astrometric quality of the data to be good enough for most scientific explorations. However, the data are not without problems and users are referred to the list of known issues before exploiting the data for science. All the data and documentations can be found at the data release site, https://hscla.mtk.nao.ac.jp/.

Vasco M. J. Henriques, Shahin Jafarzadeh, Juan Camilo Guevara Gómez et al.

In December 2016, the Atacama Large Millimeter/submillimeter Array (ALMA) carried out the first regular observations of the Sun. These early observations and the reduction of the respective data posed a challenge due to the novelty and complexity of observing the Sun with ALMA. The difficulties with producing science-ready time-resolved imaging products in a format familiar and usable by solar physicists based on the measurement sets delivered by ALMA had so far limited the availability of such data. With the development of the Solar ALMA Pipeline (SoAP), it has now become possible to routinely reduce such data sets. As a result, a growing number of science-ready solar ALMA datasets is now offered in the form of Solar ALMA Science Archive (SALSA). So far, SALSA contains primarily time series of single-pointing interferometric images at cadences of one or two seconds. The data arrays are provided in FITS format. We also present the first version of a standardised header format that accommodates future expansions and fits within the scope of other standards including the ALMA Science Archive itself and SOLARNET. The headers also include information designed to aid the reproduction of the imaging products from the raw data. Links to co-observations, if available, with a focus on those of the Interface Region Imaging Spectrograph (IRIS), are also provided. SALSA is accompanied by the Solar ALMA Library of Auxiliary Tools (SALAT) that contains IDL and Python routines for convenient loading and quick-look analysis of SALSA data.

Fernanda Roma Sobreira, Melina de Brito dos Santos, Jeorgina Gentil Rodrigues

Os estudos sobre arquivos escolares atualmente têm adquirido relevância no campo da história da educação. O objetivo deste artigo é dissertar sobre a importância dos dados arquivísticos para construção de um acervo de memória escolar. A pesquisa tem abordagem de caráter qualitativa e interpretativa, realizada na forma de estudo de caso e análise documental do arquivo escolar do Colégio Cruzeiro, fundado em 1862, localizado no Rio de Janeiro. Palavras-chave: arquivo escolar; dados escolares; cultura alemã.