Our Work

These materials are available for study purposes only and by using them you are agreeing to these specific terms of use. This short lecture provides an overview of research into health behaviour facto…

These materials are available for study purposes only and by using them you are agreeing to these specific terms of use. This short lecture provides an overview of research into health behaviour factors, including determinants of health, and how an understanding of health behaviours can help inform health investment. Originally delivered as part of an ‘introduction to health economics’ workshop for Overseas Development Institute (ODI) Health Fellows, delivered at the Centre for Health Economics at the University of York (UK) on 31 July 2019.

Benchmark of metagenomic genome binners (starting with COMEBin v1.1.0) with a repeatable harness: per-run parameters, datasets, wall time, and bin quality scored with CheckM2 and CheckM (v1). Reposito…

Benchmark of metagenomic genome binners (starting with COMEBin v1.1.0) with a repeatable harness: per-run parameters, datasets, wall time, and bin quality scored with CheckM2 and CheckM (v1). Repository: https://github.com/paulzierep/metagenomic-binning-benchmark (setup, datasets, evaluation and run docs under docs/; resume state in PROGRESS.md). This record accumulates ALL newly generated benchmark data that is not publicly available otherwise, as continuous versions under the concept DOI 10.5281/zenodo.22935024 (each new evaluated run publishes a new version). Version 4.0 (this version) contains the benchmark results of every evaluated run: baseline_rerun_autorestart1 (60 bins) cami3_v11_20260926_rerun (492 bins) fix_v11_20260924 (71 bins) human_v11_20260925 (107 bins) marine_v11_20260925 (152 bins) medium_v11_20260924 (16 bins) small_test_v4 (3 bins) sweep_001_ref (17 bins) tiny_test_n101 (1 bins). Each run contributes its bin FASTAs (runs//comebin_res/comebin_res_bins/), per-bin and aggregate CheckM2 + CheckM v1 metrics (results/.csv, runs//per_bin_results.csv), the evaluation reports (runs//eval/), run parameters and logs (run_meta.txt, comebin_run.log), all comparison figures (results/figures/, incl. the combined completeness bar plot), plus dataset, environment, fix, evaluation and preservation documentation (docs/). manifest.json inside the archive lists every file with size and SHA-256. Version 1.0 additionally contains the small functional-test dataset (comebin_small_release_v1.tar.gz, 300 contigs + overlapping reads). How to run: clone the repository, follow docs/02/06 to build the environment, run a dataset with scripts/run_comebin_fix.sh and evaluate with scripts/run_eval.sh; docs/01-datasets.md and docs/11-run-results.md describe each dataset and run.

iot-ann-ids/├── README.md├── LICENSE├── requirements.txt├── config/│   ├── model_config.yaml│   └── baselines_config.yaml├── data/│   ├── label_mapping.csv│   └── feature_list.csv├…

iot-ann-ids/├── README.md├── LICENSE├── requirements.txt├── config/│   ├── model_config.yaml│   └── baselines_config.yaml├── data/│   ├── label_mapping.csv│   └── feature_list.csv├── src/│   ├── __init__.py│   ├── config.py│   ├── preprocess.py│   ├── features.py│   ├── train_ann.py│   ├── train_baselines.py│   ├── evaluate.py│   ├── statistical_test.py│   ├── feature_importance.py│   ├── ablation.py│   └── export_onnx.py└── scripts/    └── run_all.sh ----------------------------------------------------------------------------- % MATLAB reference implementation (Deep Learning Toolbox, R2021b+)% Reproduces the preprocessing pipeline, engineered features, ANN% architecture (42 -> 256 -> 128 -> 64 -> 32 -> 5), and evaluation% protocol described in the manuscript (Sections 3.1-3.7, 4.1-4.6).%% NOTE: The reported results in the paper were obtained with the% Python/TensorFlow pipeline (src/). This file is a supplementary% reference implementation for readers preferring MATLAB.%% Author: Abdulla Juwaied% Zenodo: https://doi.org/10.5281/zenodo.23017869%      ------------------------------------------------------------------------- Copyright (c) 2026 Abdulla Juwaied Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.    

A configurable and reproducible benchmark for developing and evaluating autonomous systems in underwater gate-racing tasks, built on HoloOcean with a BlueROV2-class vehicle. Controllers operate exclus…

A configurable and reproducible benchmark for developing and evaluating autonomous systems in underwater gate-racing tasks, built on HoloOcean with a BlueROV2-class vehicle. Controllers operate exclusively from onboard observations, acoustic measurements and optional inter-vehicle communication, while an independent referee accesses privileged simulator state only to validate ordered gate crossings, enforce the rules and compute the official results.

Cet ouvrage, en libre accès et destiné aux étudiants de master, de doctorat et aux chercheurs, présente la densité de charge électronique comme l'observable c…

Cet ouvrage, en libre accès et destiné aux étudiants de master, de doctorat et aux chercheurs, présente la densité de charge électronique comme l'observable centrale de la liaison chimique dans les semiconducteurs tétraédriques et leurs alliages. Il rassemble et réorganise un programme de recherche mené depuis la fin des années 1980, pour l'essentiel à l'Université Djillali Liabès de Sidi Bel Abbès (plus de 130 articles publiés entre 1989 et 1999). L'ouvrage est construit autour de trois approches complémentaires : le calcul, par la méthode du pseudopotentiel empirique (facteurs de forme symétriques et antisymétriques, densités de valence et de bord de bande, approximation du cristal virtuel et désordre) ; l'interprétation, par les gaps homopolaire et hétéropolaire de Phillips et Van Vechten, l'ionicité et la stabilité structurale, y compris les modèles d'ionicité et de module de compressibilité développés par l'auteur et son équipe ; la mesure, par l'annihilation du positon dans le modèle des particules indépendantes (IPM) : distribution du positon, densité de moment électron-positon, repliement Lock-Crisp-West (LCW), affinité et travail de sortie. La liaison est ensuite suivie sous pression, dans les alliages désordonnés, en fonction de la température et sous contrainte. Un dernier chapitre confronte ces résultats aux méthodes ab initio modernes : DFT, DFT à deux composantes, GW, structures SQS. L'ouvrage comprend 16 figures, dont des structures de bandes, des cartes de densité de charge et du positon et un repliement LCW recalculés pour le livre, ainsi que la bibliographie complète de l'auteur pour la période 1989-1999.

Anonymous replication package for a full paper submitted to the Research Track of the Brazilian Symposium on Information Systems (SBSI 2027). Author and institutional identifiers were removed for doub…

Anonymous replication package for a full paper submitted to the Research Track of the Brazilian Symposium on Information Systems (SBSI 2027). Author and institutional identifiers were removed for double-anonymous review and will be restored after the review process. The paper presents a three-trust-zone architecture for inspecting how an ICU bed-regulation indicator is produced, preserving the distinction among observed, rule-derived, and preserved values and the effects of later revisions. On 23,980 regulation episodes, a category-based rule produced 7,285 positives and a rule that preserved existing entries produced 7,290; the net difference of five concealed 31 discordant episodes (18 and 13 in opposite directions). See README.md for the folder layout and the commands to run the tests, the fixture audit and the provenance query.