聚变科学与工程资源库
检索物理代码、实验数据、代理模型、工程几何、技术文档与验证资料，并核对来源、版本和使用条件。
BiliCore: An Open-Source Framework for LLM-Powered Applications (IRIS / AETHER / AEGIS)
Metropolitan State University of Denver, Department of Computer SciencesBiliCore is an open-source, MIT-licensed framework for building and testing LLM-powered applications. It comprises three components: IRIS (Interactive Reasoning and Integration Services) for single-agent orchestration across 100+ models from 18 provider types; AETHER (Agent Ecosystems for Testing, Hardening, Evaluation, and Research) for declarative multi-agent system creation via YAML; and AEGIS (Adversarial Evaluation and Guarding of Intelligent Systems) for security testing of multi-agent systems. Originally developed for the Colorado Sustainability Hub at the Metropolitan State University of Denver Community-Centered Computing (C3) Lab.
ginca03/Blob-dynamics-simulation
ginca03This repository contains research on blob dynamics in the Scrape-Off Layer (SOL) of tokamaks. It focuses on modeling plasma filament motion, transport properties, and interactions. The project includes data analysis, numerical simulations, and theoretical insights relevant to fusion energy.
volodymyr-yelisieiev/gk-latent-surrogate-jax
volodymyr-yelisieievJAX/Flax latent surrogate pipeline for gyrokinetic plasma turbulence snapshots
emileegraphic698/VisionFusion_OCR_QR
emileegraphic698🧠 Unlock data insights with VisionFusion, combining dynamic OCR, advanced QR decoding, and smart web enrichment for seamless integration and analysis.
Academic Observatory Workflows
Centre for Culture and Technology, Curtin UniversityAcademic Observatory Workflows provides Apache Airflow workflows for fetching, processing and analysing data about academic institutions. The workflows include: Crossref Events, Crossref Fundref, Crossref Metadata, Geonames, OpenAlex, Open Citations, ORCID, PubMed, ROR, Scopus, Unpaywall and Web of Science. Contributors Conceptualization: Richard Hosking, James P. Diprose, Aniek Roelofs, Lucy Montgomery, and Cameron Neylon. Data curation: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Cameron Neylon, Keegan R. Smith, and Julian Tonti-Filippini. Formal analysis: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Bianca Kramer, and Cameron Neylon. Funding acquisition: Lucy Montgomery and Cameron Neylon. Investigation: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Alex Massen-Hane, Keegan R. Smith, and Julian Tonti-Filippini. Methodology: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Alex Massen-Hane, Keegan R. Smith, and Julian Tonti-Filippini. Project administration: Richard Hosking, Kathryn Napier, Lucy Montgomery, and Cameron Neylon. Resources: Richard Hosking, James P. Diprose, and Aniek Roelofs. Software: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Cameron Neylon, Alex Massen-Hane, and Keegan R. Smith. Supervision: Richard Hosking, James P. Diprose, Kathryn Napier, and Cameron Neylon. Validation: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, Bianca Kramer, and Cameron Neylon. Visualization: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, Rebecca N. Handcock, and Cameron Neylon. Writing - original draft: Richard Hosking, James P. Diprose, Aniek Roelofs, Tuan-Yow Chien, and Cameron Neylon. Writing - review & editing: Richard Hosking, Aniek Roelofs, Bianca Kramer, and Cameron Neylon.
Zoverions/Learned-Multi-LoRA-Fusion
ZoverionsReact explorer for learned fusion of LoRA experts and mixture-of-adapter concepts.
Limbless-cigarbutt99/comparative-retail-sales-forecasting
Limbless-cigarbutt99Compare feature-engineered XGBoost and Temporal Fusion Transformer models to forecast multi-store retail sales.
Lauraineabsurd944/Light-WAM
Lauraineabsurd944Train lightweight world action models for robot manipulation using state-fusion decoding and frozen video backbones on LIBERO and RoboTwin2.0.
api-evangelist/niricson-software
api-evangelistNiricson Software Inc. is an infrastructure technology company making concrete infrastructure safer through AI-powered digital condition assessment. Its patented AUTOSPEX platform uses machine learning and multi-sensor data fusion (visual, thermal, and acoustic) to automatically detect, quantify, and track defects such as cracks, spalling...
api-evangelist/alif
api-evangelistAlif Semiconductor is a Silicon Valley fabless semiconductor company building secure, ultra-low-power 32-bit microcontrollers and fusion processors for edge AI/ML. Its Ensemble, Ensemble Gen AI, and Balletto product families pair Arm Cortex-M and Cortex-A cores with dedicated Ethos-U55 neural processing units, hardware AI acceleration, wireless...
api-evangelist/segments-ai
api-evangelistSegments.ai is a data-labeling platform for computer vision, supporting 2D image segmentation and vectors as well as 3D point cloud and multi-sensor fusion annotation. Its REST API and Python SDK let teams manage datasets, samples, labels, labelsets, and versioned releases programmatically for building training data pipelines.
shrvan30/flash-attention-cuda
shrvan30FlashAttention-style CUDA implementation with shared-memory tiling, online softmax fusion, IO-aware optimization, and GPU benchmarking.
hussin2323332/slrm-lumin-fusion
hussin2323332🔍 Compress datasets efficiently with Lumin Fusion, a logical engine that models high-dimensional regression using geometric sector decomposition.
PINN-ECS-HAB: Physics-Informed Neural Network with Strong PDE Residuals and a Learnable Cyst-Germination Source for Forecasting Harmful Algal Blooms in the East China Sea (code )
Zhejiang Ocean Universityv1.3.0 (2026-08-13): resubmission to Environmental Modelling & Software. Adds scripts/interpretability_analysis.py, a post-hoc interpretability diagnostic that (i) quantifies the consistency of the prescribed cyst beds with the benthic-survey envelope and the SST shelf front (2.48x front gradient over ocean background; beds within 78 km of the front), and (ii) evaluates the transport-consistency residual of the released PINN-step3 vs PINN-nophys mean fields with a single finite-difference operator (PINN-step3 residual RMS 0.83 vs 1.29 mg m-3 d-1, 1.56x smaller, with the gap largest where MODIS coverage is thinnest), outputs processed/pinn_step3/interpretability_{metrics.json,pde_residual.png,cyst_vs_benthic.png}. README, CITATION.cff and ZENODO_GUIDE updated to the EMS resubmission. All headline numbers and data are unchanged from v1.2.1; this release only adds the interpretability artifacts required by the new manuscript figures.
stellaGK/stella
stellaGKstella is a flux tube gyrokinetic code for micro-stability and turbulence simulations of strongly magnetised plasma.
api-evangelist/drover-ai
api-evangelistDrover AI is a Los Angeles-based micromobility technology company that pioneers the use of computer vision, machine learning, and proprietary sensor fusion in shared light electric vehicles (LEVs) such as e-scooters and e-bikes.
visionpilot-project/VisionPilot
visionpilot-projectOpen-source modular autonomous driving simulation platform with computer vision, deep learning, and sensor fusion. Features lane detection, object recognition, various safety features, and adaptive control in BeamNG.tech, with real-time visualization via Foxglove.
ml-jku/neural-gyrokinetics
ml-jkuMachine learning tools to accelerate high-dimensional plasma turbulence simulations.
uwplasma/GKX
uwplasmaGyrokinetic Spectral Solver of Plasma Turbulence
shrdgn/openfusion
shrdgnCombine the results from a panel of models into an enhanced response
Youjunhu/TEK
Youjunhugyrokinetic pic simulation of tokamak plasma
EmadUSyed/Plasma-Instability-in-Fusion-Simulations
EmadUSyedMachine Learning Project: Forecasting Plasma Instability in Fusion Simulations
Zulqarnain-10/Advanced-Material-Classification-Using-Sensor-Fusion-Machine-Learning
Zulqarnain-10CNN + multi-sensor fusion that sorts metal, plastic & glass for recycling - TensorFlow/Keras vision model on a Raspberry Pi + Arduino rig (~96% accuracy). Final-year capstone.
Carltonromansh946/daVinci-LLM
Carltonromansh946Build open pretraining LLMs from scratch with released data, training code, ablations, and results to improve model quality efficiently