Find open source projects actively accepting contributors. Search repositories, filter by program milestones, difficulty tags, or tech stack.
Use our Orbit AI Matcher to find out! Get instant matching scores based on your developer skills, preferred frameworks, and contribution experience.
Convert your selected open-source project into a winning GSoC, LFX, or Outreachy application using Proposal Studio.
<p>A tool for producing interactive visualizations from data generated by the <a href="https://github.com/moja-global/About_moja_global/wiki/GCBM-is-runs-CBM-science-models-on-the-FLINT-platform" target="_blank">Generic Carbon Budget Model</a> (GCBM).</p> <p>For the <a href="https://moja.global" target="_blank">moja global</a> sub-org, this project is an experiment that aims to explore the dynamic ways in which GCBM output can be viewed.</p>
<p>PODIO is a C++ library that allows the creation of event data models and efficient I/O code for HEP experiments. It does so by avoiding deep-object hierarchies and virtual inheritance. On the other hand, HDF5 is a data format that allows one to manage extremely large and complex data collections. Due to its versatility as a data model and rich set of performance features, it is an ideal format to store PODIO data in. The aim of this project is to implement an HDF5 backend for PODIO.</p>
<p>This proposal will present a scalable attributed based access control mechanism based on the Hyperledger Fabric platform. The ABAC model is a new approach to access control and supports mandatory and autonomous access control requirements based on subject, object and environment attributes. ABAC model satisfies the flexibility, dynamics, and diversity of resource management. Hyperledger Fabric supports PBFT consensus mechanism and it’s permission-ed in nature and more scalable than public blockchains.</p>
<p>The project aims at improving the audio track analysis features to compute information such as downbeat, time signature, measures, and phrases.</p> <p>Downbeat detection will pave the way for bar detection and time signature computation. This feature is necessary since a song can be in a time signature other than 4/4, which will cause most beat and time-related features such as looping and quantization inconsistent with the track. Overall smarter quantization can be achieved with the added information.</p>
<p>Toolkit for Multivariate Analysis (TMVA) is a multi-purpose machine learning toolkit integrated into the ROOT scientific software framework, used in many particle physics data analysis and applications. Currently, we are developing a fast inference system in TMVA that takes the ONNX model as input and produces compilation-ready standalone C++ scripts as output. These scripts will then provide users an easy way to deploy their deep learning models in their physics software and analysis frameworks.</p>
<p>Most of the data from IACTs currently in operation comes in ROOT format. Enabling ROOT input in CTLearn will open its doors to existing IACT facilities. Hence, this project aims to use CTLearn to train classification models that tell gamma-ray images from cosmic-ray images apart and why this is important in the context of IACTs, then, enable the input of data in ROOT format and train some of the already existing deep learning models in CTLearn with data from current-generation IACTs.</p>
Symmetry is one of the most beautiful and interesting phenomena in physics. Particle Physics is dominated by Lorentz symmetry. It is seen that Equivariant networks are more expressive and sample efficient than conventional Neural Networks. The project aims to build a Lorentz Equivariant Neural network to analyse particle collision data obtained from CMS Open Data. The project also aims to compare the results of the Lorentz Equivariant model with other Equivariant models(symmetry groups) and conventional CNNs and ViTs in terms of performance, speed and sample efficiency.
<p>This project when completed gives ardupilot the ability to control a walking robot. To be able to do that ardupilot needs new modules which support the kinematics model required by the robot . The walking robot will also have support for existing modes and new vehicle specific modes.</p> <ol> <li>Implementation of walking robot over ardupilot </li> <li>Solid kinematics model </li> <li>Simulation support </li> <li>GCS support </li> <li>Documentation</li> </ol>
<p>My goal for GSoC 2020 is to implement, test, and maintain a higher-level API for Gaussian Processes in PyMC4 using TensorFlow and TensorFlow Probability and write tutorials/articles and notebooks explaining their usage.</p> <p>My work consists of implementing Gaussian Process Modes and writing optimization methods like Newton's Method and Powell's Method to find the maximum a-posteriori of the models. My goal is also to implement at least one approximation technique when full GP modelling becomes impractical.</p>
<p>ArviZ is a Python package for exploratory analysis of Bayesian models, intending to provide backend-agnostic tools for diagnostics and visualizations of Bayesian inference in Python. There are currently no specific plots in arviz to visualize time series or regression models. My proposal is to add time series plots including line and scatter plots to the library to visualize the uncertainty in the actual data vs the range of predicted data. Dataset is assumed to be provided by the user. The project includes the documentation of functions implemented and examples for each of them.</p>
For this project, I plan to explore state space models(SSM) and variational inference(VI) in depth. First, I will translate legacy numpy/PyTorch codes into JAX/Flax and contribute additional codes and figures of pedagogical utility to the relevant chapters in the PRML textbook. Then, I will expand the JAX SSM Library(JSL), with a primary focus on implementing Gaussian smoothers and particle smoothers. Last, I will explore some of the state-of-the-art particle-based variational inference models and implement them to add to the pyprobml repository.
The proposed implementation is planned to include building the camera-based warning system, starting from the low-level driver implementation, creating necessary logic inside PipeWire nodes, training AI models for objects detection, and designing the suitable User Interface using Flutter. The final product would be a Warning System that leverages implemented inference models (e.g. Object Detection) and accordingly issues a visual alert in response to critical conditions, such as the detection of an object in the blind spot, or a very low time-to-collision.
A key component of Bayesian inference is integrating over prior distributions to obtain posteriors. In practice however, these distributions are often high-dimensional, resulting in a significant computational cost associated with integration, which remains a key challenge in Bayesian ML. Under certain assumptions, it is possible to efficiently compute posteriors for Latent Gaussian Models (LGMs), which represent a broad class of statistical models in Bayesian statistics. This is known as the method of Integrated Nested Laplace Approximations (INLA), and this project aims to implement a working basis for INLA in the PyMC library.
<p>This proposal aims to create a ROS package which will allow any arbitrary robotic manipulator to intelligently explore and map it’s work cell environment by utilizing a 3-D sensor based perception pipeline. This will provide a smart alternative to the current method where the work cell environment is defined using manually created URDF models. The 3-D model of the environment provided as an output can be utilized by packages such as MoveIt! for path planning and collision detection for the specific task at hand.</p>
gprMax is CLI interface software which is used to solve electrodynamics problems by solving Maxwell's Equations using FDTD method. It is also used to simulate Ground Penetrating Radar for various industry and academia purpose. But it's installation can become a bit cumbersome and nowadays people prefer a GUI interface. Hence I helped to develop a custom gprMax kernel which can be used to build and run models in Jupyter Notebooks. I also developed a python function that can be used to visualize the 3D geometry of the models in notebook environment.
This project investigates the potential of quantum machine learning methods, such as quantum feature encoding and quantum neural networks (QNNs), for the characterization of exoplanet atmospheres using transmission spectra. The goal is to assess how well quantum representations preserve and separate atmospheric features compared to classical approaches, and to determine whether QML-based models can provide meaningful advantages in estimating atmospheric properties and identifying molecular signatures. Performance is benchmarked against established classical models to evaluate both the scientific utility and the practical feasibility of QML in this domain.
<p>The project proposed aims to implement an <strong>Arabic speech recognition</strong> model using training data from the <a href="http://www.mgb-challenge.org/arabic.html" target="_blank"><em>MGB-3</em></a> Arabic datasets to perform speech recognition on a <strong>television news corpus</strong> captured in Cairo by Red Hen Lab. Another contribution will be to train a <strong>dialect identification</strong> model to perform Arabic dialect identification on the Television news corpus.</p>
This project builds a fully explainable NLP pipeline to identify and classify Military-Involved Conflicts (MIC) from raw text articles. Unlike pretrained models, it uses a custom heuristic approach—extracting countries, actions, dates, fatalities, and sentiment based on rule-based logic. Though this results in longer processing times, it ensures complete transparency, flexibility, and interpretability, which is critical for social science and humanitarian research. The pipeline can later be scaled using models like BERT to enhance classification while retaining its explainable foundation.
<p>This project is roughly divided into three parts: OCR Recognition, which uses existing tools to extract captions from videos to text; Speech Recognition, which uses deep learning tools(BaiduSpeech) to translate audios to text; NLP tasks, including segmentation, part-of-speech tagging, named entity recognition, dependency parsing, semantic role labeling and so on. The most important part is Speech Recognition. Since there are few guidences about how to use DeepSpeech model to train Chinese, I will pay more attention to this part and train a model as soon as possible.</p>
<p>The objective of this project is to automatically detect types of accountability in news articles when describing crimes such as shootings, using a dataset that was annotated with custom labels for this task. Analysis will be done to assess the given annotations, and build a classifier to automate this type of annotation for future research. Experiments will be conducted to compare various text representations and classification models. AutoML approaches will be used to tune parameters and optimize models. This project is in collaboration with Red Hen Lab, Dr. Glik from the UCLA school of public health.</p>
While the OpenAPI specifications for Kubernetes offer a taxonomy, integrating a graph data model with formalized ontologies unlocks a multitude of capabilities. Among these, enabling inferencing necessary for natural language processing stands out as a straightforward application. This, in turn, facilitates the possibility of a human-centric query/response interaction. Importantly, advancing to a knowledge semantic graph from a connected systems' graph data model opens the door to building more sophisticated systems. Expected Outcome: - Identifying new technologies from CNCF landscape and creating new YAML-formatted definitions of one or more relationships.
Time series forecasting is paramount in many domains, including finance, healthcare, energy, and climate science. This project suggests incorporating deep learning-based forecasting models—Informer, TCN, and DeepAR—into the aeon/tookit. The objectives are to construct an efficient and scalable framework for forecasting that accommodates top-performing models, is compatible with Aeon’s data management, and provides stable training, evaluation, and documentation. Through simplifying the availability of advanced forecasting software, the project would make it easier to utilize the toolkit to assist researchers with streamlined time series analysis.
The goal of this project is to improve currently existing implementations for rounding convex polytopes from Volesti as well as develop a new method and analyze its performance. The new method be using the available C++ implementation from Volesti of the Khachiyan algorithm for solving the minimum volume enclosing ellipsoid (MVEE), which will then be transformed into the maximum volume ellipsoid of the polytope (MVE). Finally, the polytope will be brought into John position using the transformation that maps the MVE to the unit ball. The performance of the improved implementations and the new one will then be compared to existing rounding procedures from PolyRound.
Open Climate Fix recent Cloudcasting model is able to predict clouds movements very accurately but for end user it’s hard to see or visualize the results, this project aims to create a new dashboard similar to quartz-frontend that is already up and running, to visualize results from the cloudcasting model the dashboard will be build using Next.js and MapBox as base map layer, Sentry to log errors, and a minimal fastapi server written in python to preprocess zaar data and serve it to the frontend, this solution will be a proof of concept for the new MVP