Whats new?
This portal provides a new entry for the repositories and documentation of this project.
Welcome to ML-PPA
The Machine Learning-based Pipeline for Pulsar Analysis (ML-PPA) software framework for extracting pulsar signals in data streams from radio astronomical antennas.
Pipelines and Notebooks
The software framework provides Python and C++ written pipelines. Most of the pipeline can easily be analyzed using jupyter notebooks. The following sub-projects are currently available.
All projects provide a link to a repository and some of them also a link to its documentation.
PulsarSA
Tools to segment and analyse pulsar signals from noisy frequency-time data.
PulsarDT
Simulation of pulsars with customisable emissions to generate training data for AI detection.
PulsarDT++
A C++ version of PulsarDT.
Telescope Noise DT
Simulation of radio telescope noise based on modular libraries.
PulsarRFI_NN
A wrapper around PulsarRFI_Gen, creating a simulated data pipeline for training and testing.
PulsarRFI_Gen
A tool for generating synthetic radio pulses as well as primitive broadband and narrowband RFIs.
DADA Noise Analysis
A Jupyter notebook for reading and analysing data from the DADA radio telescope.
Publications
Discover research papers, reports, posters, presentations and publications related to this project.
Dynamization of big data workflows
A presentation at the ADASS 2025 conference in Görlitz, Germany (Heßling et al., 2025).
Multi-messaging: How can uniform access to astronomical archives be achieved?
A Bird of a Feather Session at the ADASS 2025 conference in Görlitz, Germany (Heßling et al., 2025).
Software framework for Pulsar Detection using Machine Learning and Digital Twin
A poster at the ADASS 2025 conference in Görlitz, Germany (Saha et al., 2025).
DZA HPC-Cloud
A poster at the ADASS 2025 conference in Görlitz, Germany (Haupt et al., 2025).
Memory-based Computing: First Results
A poster at the ADASS 2025 conference in Görlitz, Germany (Buchholz et al., 2025).
ML-PPA
A document explaining the ML-PPA software framework (Kazantsev et al., 2025).