Our Work

Version 1.7 (2026-09-23): the review manuscript as submitted to the IEEE Transactions on Information Theory, a snapshot of paper/tit-cr-context at tag tit-cr-context-1.7 of github.com/ahb-sjsu/geometr…

Version 1.7 (2026-09-23): the review manuscript as submitted to the IEEE Transactions on Information Theory, a snapshot of paper/tit-cr-context at tag tit-cr-context-1.7 of github.com/ahb-sjsu/geometric-observation. Single-column review format, 33 pages; figures set in Type 1 fonts through the pgf backend; the AI-assistance acknowledgment names the systems, sections and level of assistance. Title and mathematics unchanged from 1.5. Files: the manuscript PDF, the executed reproduction notebook (reproduce.ipynb), and a zip of the manuscript directory with its LaTeX source, figures, figure scripts, independent numerical and symbolic verification scripts, MATLAB checks, and archived earlier versions.An encoder observes a jointly Gaussian pair (Y, V) and describes Y for a decoder that sees the description alone; a third party retains a noisy copy S = V + U of the context and never decodes. The paper evaluates Steinberg's common-reconstruction functional on this augmented source: it gives the minimal conditional content in closed form, the exact rate-content region with its two-noise-fraction Pareto frontier and a strict tradeoff exactly when the described variable and the context are partially correlated, non-determination by the reduced (Y, S) law, the vector-context frontier, the binary rate-content frontier through a tilt equation, and a direct Gaussian operational theorem.

Fondé sur 24 entretiens et six observations menés en 2026 dans six structures du district sanitaire de Sikasso (Mali), cet article propose une lecture anthropologique du tri des d&eacute…

Fondé sur 24 entretiens et six observations menés en 2026 dans six structures du district sanitaire de Sikasso (Mali), cet article propose une lecture anthropologique du tri des déchets biomédicaux à partir de trois modèles théoriques : la souillure douglassienne, la liminalité turnérienne et le regard disciplinaire foucaldien. Le code couleur, bien que connu des acteurs, est fréquemment transgressé : ruptures de boîtes de sécurité, recapuchonnage, mélange des catégories de déchets. Distinguant déclarations, observations et interprétation théorique, l’article montre que ces écarts relèvent moins d’une négligence technique que d’un arbitrage entre un ordre symbolique partagé et une infrastructure matérielle déficiente.

Ce Projet de Fin d’Études, mené chez Thales SIX GTS France, vise l’industrialisation du processus de génération des composants logiciels du produit SSR. Face &a…

Ce Projet de Fin d’Études, mené chez Thales SIX GTS France, vise l’industrialisation du processus de génération des composants logiciels du produit SSR. Face à un existant hétérogène et peu automatisé, limitant reproductibilité, traçabilité et maintenabilité, le projet propose une chaîne de génération logicielle industrialisée, intégrant les principes DevSecOps et les bonnes pratiques modernes d’ingénierie logicielle, tout en respectant les contraintes spécifiques de Thales. Après une analyse de l’existant et une étude des outils industriels (CMake, Jenkins, Artifactory), une architecture cible de pipeline a été conçue et implémentée sous forme de démonstrateur fonctionnel. Un Proof of Concept a validé la chaîne complète, démontrant des gains significatifs en reproductibilité, standardisation et traçabilité des builds, ouvrant la voie à une industrialisation opérationnelle.

PCAP2Flows v1.0 PCAP2Flows is an interactive tool for converting network packet captures (.pcap) into flow-based CSV datasets using multiple flow-generation engines within a reproducible Docker enviro…

PCAP2Flows v1.0 PCAP2Flows is an interactive tool for converting network packet captures (.pcap) into flow-based CSV datasets using multiple flow-generation engines within a reproducible Docker environment. It is designed for researchers, students, and practitioners working with network traffic data who want to generate, compare, and process flow-based datasets without manually installing and configuring several independent tools. Main features PCAP2Flows supports three flow-generation engines: Argus NTLFlowLyzer Zeek with zeek-flowmeter It also provides post-processing functions for: filtering flows by port removing constant columns adding labels derived from filenames generating flow-file statistics Intended use PCAP2Flows is intended for: network traffic analysis cybersecurity research intrusion-detection dataset generation IoT traffic analysis flow-based machine-learning datasets reproducible network measurement comparison of flow-generation tools academic research and education Reproducible Docker-based release PCAP2Flows v1.0 is distributed as a complete Docker-based release. The supplied Docker image contains the tested execution environment together with the required third-party tools and dependencies, including: Argus and ra NTLFlowLyzer Zeek zeek-flowmeter Python and required Python packages operating-system dependencies Users therefore do not need to install these components separately. The Docker image also preserves the tested third-party environment, helping protect the release against future changes to package repositories, dependency versions, and upstream software. The PCAP2Flows-specific Python source code remains available separately under the source/ directory and can be modified without rebuilding the Docker image. Version   PCAP2Flows version: 1.0 Docker image: pcap2flows:1.0   Platform support The release is intended to run on: Windows using Docker Desktop with WSL 2 macOS using Docker Desktop Linux using Docker Engine and Docker Compose Separate installation and usage guides are provided for each platform. Release contents The PCAP2Flows release includes:   PCAP2Flows-release/ ├── config/ ├── docker-compose.yml ├── input_pcaps/ ├── output/ ├── pcap2flows-1.0.tar ├── pcap2flows-1.0.tar.sha256 └── source/   The main directories are:   input_pcaps/ User-provided PCAP files output/ Generated flow datasets and processing results config/ Persistent runtime configuration source/ Editable PCAP2Flows Python source code   The file:   pcap2flows-1.0.tar   contains the complete Docker image for the PCAP2Flows v1.0 execution environment. Citation If PCAP2Flows is used in published research, please cite the associated publication and Zenodo software record. Suggested citation format:   Author(s). PCAP2Flows v1.0: A reproducible multi-engine tool for converting packet captures into network-flow datasets. Zenodo. Version 1.0. DOI: [10.5281/zenodo.22116697]   License PCAP2Flows includes and/or depends on several third-party open-source components. Each component remains governed by its own license: Argus — GNU General Public License v3.0 (GPL-3.0). NTLFlowLyzer — GNU General Public License v3.0 (GPL-3.0). Zeek — BSD license. zeek-flowmeter — MIT License.   Funding This work has been supported by the Grant DISCOVERY (PID2023-148716OB- C33) funded by MICIU/ AEI/10.13039/501100011033 and FEDER, UE.  

Davlat ko‘magini oluvchilar uchun yagona modeldan voz kechib, biznesning rivojlanish bosqichi, mavjud cheklov turi hamda qarorlarni o‘zgartirishga tayyorlik darajasiga ko‘ra farqlanu…

Davlat ko‘magini oluvchilar uchun yagona modeldan voz kechib, biznesning rivojlanish bosqichi, mavjud cheklov turi hamda qarorlarni o‘zgartirishga tayyorlik darajasiga ko‘ra farqlanuvchi xulq-atvor segmentlari tizimiga o‘tish taklif etiladi. O‘zbekistonning rasmiy statistik ma’lumotlari kichik korxonalar tarkibining yuqori darajada dinamik ekanini, xalqaro taqqoslash esa rasmiy tadbirkorlikka kirish imkoniyatlari bo‘yicha hali foydalanilmagan salohiyat mavjudligini ko‘rsatadi. Manzilli qo‘llab-quvvatlash vositalari portfelini faqat xizmat ko‘rsatilgan arizachilar soni bo‘yicha emas, balki firmalarning rivojlanish bosqichlari o‘rtasidagi o‘tishlari va ularning xulq-atvor barqarorligi asosida baholash lozim.

CONTENTS Cha. No. Chapters Page No. - PREFACE 01 1 WIT AND WISDOM 01-39 2 VALUES AND DISPOSITIONS 01-37 3 MYSTERY AND MAGIC 01-39 4 ENVIRONMENT 01-30…

CONTENTS Cha. No. Chapters Page No. - PREFACE 01 1 WIT AND WISDOM 01-39 2 VALUES AND DISPOSITIONS 01-37 3 MYSTERY AND MAGIC 01-39 4 ENVIRONMENT 01-30 5 SCIENCE AND CURIOSITY 01-40

This dataset accompanies the paper "Inferring Gender from Author Names with Local LLMs: A Multi-Model Evaluation" (Journal of Informetrics, 2026). It contains 194,528 first name–country pairs de…

This dataset accompanies the paper "Inferring Gender from Author Names with Local LLMs: A Multi-Model Evaluation" (Journal of Informetrics, 2026). It contains 194,528 first name–country pairs derived from Wikidata, along with gender predictions (Man / Woman / Unknown) produced by 25 open-weight large language models run locally via Ollama, ranging from 270M to 70B parameters. The ground truth is based on the gender distribution of real people registered in Wikidata for each name and country of origin.  The dataset includes: (1) the ground-truth reference file with name, country, and sex; (2) the Wikidata source file with gender probability and frequency statistics; (3) one prediction file per model (25 total); and (4) Python scripts to reproduce the evaluation metrics (precision, recall, F1-score) and all figures reported in the paper.  Data are released under CC BY 4.0; code under Apache 2.0

A utilização de plataformas digitais no Exame Clínico Objetivo Estruturado (OSCE) tem ampliado a necessidade de compreender seus impactos na experiência dos docentes avaliad…

A utilização de plataformas digitais no Exame Clínico Objetivo Estruturado (OSCE) tem ampliado a necessidade de compreender seus impactos na experiência dos docentes avaliadores. Este estudo analisou a associação entre fatores estruturais da arquitetura digital do OSCE, incluindo clareza da estrutura avaliativa, padronização dos processos, qualidade da interface e acompanhamento pedagógico, com a percepção de justiça avaliativa e indicadores contextuais de autoeficácia docente na educação médica. Trata-se de um estudo quantitativo, de delineamento correlacional, realizado com 60 docentes de uma instituição de ensino superior, dos quais 21 participaram de uma segunda etapa após um simulado padronizado do OSCE. Os dados foram coletados por meio de questionários e analisados utilizando estatística descritiva e correlação de Pearson. Os resultados evidenciaram associações positivas e significativas entre qualidade da interface, clareza dos critérios avaliativos, padronização dos processos, acompanhamento pedagógico e percepção de justiça avaliativa. As correlações mais elevadas foram observadas entre interface e clareza (r = 0,863), interface e equidade (r = 0,828) e acompanhamento e interface (r = 0,843) (p < 0,001). Conclui-se que ambientes digitais mais organizados, intuitivos e padronizados estão associados a percepções mais positivas de transparência, confiabilidade e segurança no processo avaliativo, destacando a relevância da arquitetura digital para a qualidade das avaliações na educação médica.

core library code