[{"data":1,"prerenderedAt":213},["ShallowReactive",2],{"research":3},{"_path":4,"_dir":5,"_draft":6,"_partial":6,"_locale":5,"research":7,"_id":207,"_type":208,"title":209,"_source":210,"_file":211,"_stem":212,"_extension":208},"/research","",false,[8,31,50,71,97,116,133,151,163,178,193],{"title":9,"cover":10,"paper_link":11,"topics":12,"supervision":5,"contributors":17,"status":25,"venue":26,"featured":27,"abstract":28,"links":29},"Phoenix: Safe End to End Codebase Refactoring via Multi-Agent LLMs","https://placehold.co/600x400?text=Phoenix","https://arxiv.org/abs/2606.20243",[13,14,15,16],"LLM Agents","AI Safety","Safety Gates","Evaluation",[18,21,23],{"name":19,"affiliation":20},"Kipngeno Koech","Carnegie Mellon University Africa",{"name":22,"affiliation":20},"Muhammad Adam",{"name":24,"affiliation":20},"Joao Barros","Published","International Conference on Responsible AI (ICRAI) 2026",true,"Autonomous LLM agents can synthesize code changes at scale, but deploying them safely on real codebases requires guarantees that generated patches preserve correctness. Phoenix is a multi-agent LLM system that resolves GitHub issues from triage through pull-request creation, combining seven layered safety controls with a baseline-aware test evaluation strategy. Work is decomposed across six specialized agents — planner, reproducer, coder, tester, failure analyst, and PR agent — coordinated by a label-based GitHub webhook state machine, with every change checked against a baseline test run before a pull request is opened.",{"preprint":11,"code":30,"data":30},"#",{"title":32,"cover":33,"paper_link":30,"topics":34,"supervision":39,"contributors":40,"status":46,"venue":47,"featured":6,"abstract":48,"links":49},"An Adaptive Latent Semantic Analysis Framework for Binary Text Classification","https://placehold.co/600x400?text=A-LSA+Classification",[35,36,37,38],"NLP","Latent Semantic Analysis","Interpretability","Text Classification","Prof. Kolyang",[41,44],{"name":42,"affiliation":43},"Isaac Touza","University of Maroua",{"name":45,"affiliation":43},"Kaladzavi Guidedi","Submitted","International Journal of Information Management","Binary text classification is a fundamental task in natural language processing, typically addressed using discriminative machine learning models or deep neural architectures. While Latent Semantic Analysis (LSA) has historically played an important role in semantic representation, it is now largely confined to a preprocessing function due to its static and task-independent formulation. This paper proposes an Adaptive Latent Semantic Analysis (A-LSA) framework that redefines LSA as a discriminative, class-aware classification model. The proposed approach constructs class-conditional latent semantic spaces and introduces a semantic differential distance as a direct decision criterion. Experimental results on benchmark datasets demonstrate that A-LSA achieves competitive performance compared to standard classifiers, while offering superior interpretability and significantly reduced computational cost (inference time 40x faster than BERT).",{"preprint":30,"code":30,"data":30},{"title":51,"cover":52,"paper_link":53,"topics":54,"supervision":59,"contributors":60,"status":46,"venue":67,"featured":27,"abstract":68,"links":69},"Zero-Shot Neural Priors for Generalizable Cross-Subject and Cross-Task EEG Decoding","/eeg-decoding.png","https://arxiv.org/abs/2606.23706",[55,56,57,58],"EEG","Brain-Computer Interfaces","Zero-Shot Learning","Neural Decoding","Carnegie Mellon University",[61,63,65],{"name":62,"affiliation":20},"Brandone Fonya",{"name":64,"affiliation":20},"Nchofon Tagha Ghogomu",{"name":66,"affiliation":20},"Pauline Nyaboe","Medical Imaging with Deep Learning (MIDL) 2026","The development of generalizable electroencephalography (EEG) decoding models is essential for robust brain-computer interfaces (BCI) and objective neural biomarkers in mental health. Conventional approaches have been hindered by poor cross-subject and cross-task generalization, owing to high inter-subject variability and non-stationary neural signals. We address this challenge with a zero-shot cross-subject decoding framework on the large-scale Healthy Brain Network dataset,benchmarking a convolutional neural network baseline, a hybrid LSTM, and a Transformer-based foundation model. To adapt the Transformer for regression while averting catastrophic forgetting, we propose a novel progressive unfreezing strategy. The baseline yielded an nRMSE of 0.9991, whereas our fine-tuned Transformer achieved 0.9799 on unseen subjects.This work establishes scalable, calibration-free EEG decoding for computational psychiatry and behavioral prediction.",{"preprint":53,"code":70,"data":30},"https://github.com/baimamboukar/EEG-cross-subject-decoding",{"title":72,"cover":73,"paper_link":74,"topics":75,"supervision":5,"authors":80,"status":93,"venue":94,"featured":27,"abstract":95,"links":96},"Retrieval with Multiple Query Vectors through Anomalous Pattern Detection","/retrieval.png","https://arxiv.org/abs/2605.01965v1",[76,77,78,79],"Information Retrieval","DeepScan","LLM Embeddings","Anomalous Pattern",[81,84,86,89,91],{"name":82,"affiliation":83},"Allassan Tchangmena A Nken","University of Galway",{"name":85,"affiliation":20},"Baimam Boukar Jean Jacques",{"name":87,"affiliation":88},"Miriam Rateike","IBM Research Africa",{"name":90,"affiliation":88},"Celia Cintas",{"name":92,"affiliation":88},"Skyler Speakman","Preprint","AAAI 2026 Workshop on New Frontiers in Information Retrieval","A classical vector retrieval problem typically considers a single query embedding vector as input and retrieves the most similar embedding vectors from a vector database. However, complex reasoning and retrieval tasks frequently require multiple query vectors, rather than a single one. In this work, we propose a retrieval method that considers multiple query vectors simultaneously and retrieves the most relevant vectors from the database using concepts from anomalous pattern detection. Specifically, our approach leverages a set of query vectors Q (with |Q| ≥ 1), and identifies the subset of vector dimensions within Q that standout (anomalous) from the rest of dimensions. Next, we scan the vector database to retrieve the set of vectors that are also anomalous across the previously identified vector dimensions and return them as our retrieved set of vectors. We validate our approach on two image datasets, a text dataset, and a tabular dataset.",{"preprint":74},{"title":98,"cover":99,"paper_link":30,"topics":100,"supervision":104,"contributors":105,"status":25,"venue":112,"featured":6,"abstract":113,"links":114},"Mapping Socioeconomic Air Quality Disparities In Rwanda Using Sentinel-5P TROPOMI Data In Google Earth Engine","https://raw.githubusercontent.com/baimamboukar/air-quality-socioeconomic-analysis-research/refs/heads/main/results/figures/distribution%20in%202025.png",[101,102,103],"Air Quality","Remote Sensing","Google EE"," Emily Aiken",[106,108,110],{"name":107,"affiliation":20},"Kamikazi Raissa",{"name":109,"affiliation":20},"Bertin Ndahayo",{"name":111,"affiliation":20},"Evelyne Umubyeyi","IEEE MIGARS 2025","This study investigates socioeconomic disparities in air quality across Rwanda using Sentinel-5P TROPOMI satellite data processed through Google Earth Engine. We develop novel methodologies to correlate atmospheric pollutant concentrations with demographic and economic indicators, revealing significant patterns in environmental justice. Our analysis demonstrates how satellite-based monitoring can inform policy decisions for equitable environmental health outcomes.",{"preprint":30,"code":115,"data":115},"https://github.com/baimamboukar/air-quality-socioeconomic-analysis-research",{"title":117,"cover":118,"paper_link":119,"topics":120,"supervision":124,"contributors":125,"featured":27,"status":128,"venue":129,"abstract":130,"links":131},"Explainable Deep-Learning Based Potentially Hazardous Asteroids Classification Using Graph Neural Networks","/asteroids.png","https://arxiv.org/html/2504.18605v1",[121,122,123],"Space","AI","Astrophysics"," Clarence Worrel",[126],{"name":127,"affiliation":20},"Clarence Worrel","Presented","Harvard AstroAI Workshop 2025.","Classifying Potentially Hazardous Asteroids (PHAs) is crucial for planetary defense. While the official designation relies on simple metrics (Earth MOID and Absolute Magnitude H), we explore if complex dynamical relationships offer a robust, interpretable hazard assessment framework. We introduce a Graph Neural Network (GNN) approach that models asteroids as nodes defined by 11 orbital and physical features. Edges are defined using the Tisserand parameter relative to Jupiter (TJ) as a proxy for dynamical similarity, allowing the GNN to capture latent relationships among asteroid groups. Using the NASA/JPL Small-Body Database dataset, the model achieves an overall accuracy of 99% and an AUC of 0.99.",{"preprint":119,"code":132,"data":132},"https://github.com/baimamboukar/hazardous-asteroid-classification",{"title":134,"cover":135,"paper_link":30,"topics":136,"supervision":139,"contributors":140,"status":147,"featured":27,"venue":148,"abstract":149,"links":150},"Computer Vision-based Calibration of Visual Landing Aids Using Autonomous Drones (PAPI Case Study)","/calibration.png",[137,122,138],"Computer Vision","Aerospace"," Moise Busogi",[141,143,145],{"name":142,"affiliation":20},"Alice Mugengano",{"name":144,"affiliation":20},"Jonathan Kayizzi",{"name":146,"affiliation":20},"Richard Muhirwa","Review","IEEE Aerospace Conference 2026.","This research presents a vision-based approach to autonomous drone navigation and landing trajectory optimization, demonstrated for the calibration of Precision Approach Path Indicators (PAPI). Traditional calibration methods require specialized aircraft equipped with photometric equipment, resulting in high operational costs (5,000-25,000 per calibration), runway disruptions, and significant carbon emissions. Our proposed system leverages computer vision and machine learning to detect and classify PAPI light configurations under diverse environmental conditions. A convolutional neural network, trained on the LARD dataset comprising over 17,000 real and synthetic aerial runway images, enables accurate pattern recognition. This is integrated with a trajectory generation algorithm that aligns drone flight paths to evaluate and verify standard approach angles. The system achieves a modal classification accuracy of 97% in detecting PAPI light patterns and a fivefold improvement in angular measurement precision (±0.01° vs. ±0.05°).",{"preprint":30,"code":30,"data":30},{"title":152,"cover":153,"paper_link":30,"topics":154,"supervision":139,"contributors":157,"status":25,"featured":27,"venue":112,"abstract":160,"links":161},"Humidity Inference with Geographic Features and Machine Learning for Enhanced Contrail Prediction for African Airspace","/contrails.png",[122,155,156],"Aviation","Contrails",[158,159],{"name":142,"affiliation":20},{"name":144,"affiliation":20},"Air travel's 3.8% yearly growth amplifies aviation's impact, with contrails causing about 4% of global warming. In Africa, sparse meteorological data and coarse ERA5 resolution underpredict contrails in cities like Nairobi and Lagos. This study uses a Random Forest model within PyContrails to infer upper-tropospheric humidity (RHi) from water proximity, forest density, and topography. Using global datasets, it achieves an R2 of 0.59 identifying water proximity as the main driver. African cities show only 32 contrail segments versus 450 in Europe, underscoring the model's potential to refine predictions and reduce aviation's non-CO₂ climate effects.",{"preprint":30,"code":162,"data":162},"https://github.com/JonahKayizzi/Environment-Impact-Assessment-of-Aircraft-Contrails",{"title":164,"cover":165,"paper_link":30,"topics":166,"supervision":169,"contributors":170,"status":175,"venue":5,"featured":6,"abstract":176,"links":177},"Thermography for Breast Cancer Diagnosis using AI","https://placehold.co/600x400?text=Breast+Cancer+AI",[122,167,168],"Healthcare","Medical Imaging"," Edith Luhanga",[171,173],{"name":172,"affiliation":20},"Peace Bakare",{"name":174,"affiliation":20},"Lucie Niyomutoni","Ongoing","Early detection of breast cancer is crucial for patient outcomes, particularly in resource-constrained settings. This study develops an AI-powered thermography system for breast cancer screening using deep learning approaches. Our method provides a non-invasive, cost-effective alternative to traditional screening methods, making it accessible for underserved populations.",{"preprint":30,"code":30,"data":30},{"title":179,"cover":180,"paper_link":30,"topics":181,"supervision":185,"contributors":186,"status":175,"venue":189,"featured":6,"abstract":190,"links":191},"Causal Structure Analysis for Telemetry Anomaly Detection in Spacecraft Systems","/telemetry.png",[182,183,184],"Satellites Operations","Telemetry","Graphical Models"," Assane Gueye",[187],{"name":188,"affiliation":20},"Kip Koech","10th TTC International Workshop","Spacecraft telemetry anomaly detection is critical for mission success and safety. This research applies causal inference methods to understand the underlying causal structure in spacecraft telemetry data. We developed graphical models that can identify root causes of anomalies, enabling more effective predictive maintenance and mission planning strategies.",{"preprint":30,"code":192,"data":192},"https://github.com/baimamboukar/causual-structure-discovery-spacecraft-telemetry",{"title":194,"cover":195,"paper_link":30,"topics":196,"supervision":5,"contributors":200,"status":204,"venue":205,"abstract":206},"Towards the Africa we want: What is the role of space technology in achieving the African Union Agenda 2063?","https://placehold.co/600x400?text=Space+Policy",[197,198,199],"African Union","Satellites","AfSA",[201,203],{"name":202,"affiliation":20},"Manyara Baraka",{"name":188,"affiliation":20},"In Progress","Newspace Africa Conference 2026","The African Union's Agenda 2063 outlines ambitious goals for continental transformation. This policy analysis examines how space technology can accelerate progress toward these objectives, focusing on applications in agriculture, telecommunications, climate monitoring, and governance. We provide recommendations for space policy frameworks that align with Africa's development priorities.","content:research.yaml","yaml","Research","content","research.yaml","research",1786626763531]