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.
The existing GPU profiling workflow in OpenVino, through benchmark_app connects kernels to the model layers (primitives) and reports their execution times. However, this level of abstraction is too coarse to identify kernel instruction level insights. Intel GTPin, on the other hand, provides instruction level GPU kernel metrics like clock cycle counts, but is unable to directly map such metrics to specific inference or layers. This project aims to build a GTPin-centered profiling capability within OpenVINO to identify hotspot kernels and analyze their execution characteristics. The approach involves integrating controlled GTPin profiling into the inference flow, developing a custom GTPin tool for collecting kernel-level metrics, and progressively interpreting these metrics using signals from OpenVINO performance counters, and kernel metadata. The expected deliverable is a GTPin-based profiling workflow integrated into OpenVINO, capable of identifying bottleneck kernels, providing insight into their behavior, and offering indicative optimization directions, with validation using tools such as VTune.
The std::arch in the standard library provides architecture-specific intrinsic functions, which directly map to a single machine instruction. The ultimate goal is to inline stdarch into rustc completely and archive the stdarch repository making stdarch tests reliable and maintainable inside rustc infrastructure. Which we will achieve by incrementally investigating the CI test suite of stdarch and port as much of it into rust-lang/rust. The clear set of deliverables for this project are Map stdarch CI to rustc, land x.py steps for verify + stdarch tests, enable core_arch (plain then assert_instr on Linux), port intrinsic-test in phases (host, then cross if feasible), minimal stdarch-gen checks, wire CI, document, final report.
This project builds an AI-powered agent for KnowledgeSpace using Retrieval-Augmented Generation (RAG) to deliver precise, context-aware, and scientifically accurate answers. It combines Elasticsearch results with vector search (e.g., Vertex AI Vector Search) through an agentic system implemented in LangGraph, which orchestrates tool calls and fuses evidence for high-quality responses. The backend is a FastAPI service deployed on Google Cloud Platform with Vertex AI, and the frontend is a responsive React application. An automated update script regularly ingests new datasets and refreshes indices/embeddings to keep answers current. Together, these components streamline how neuroscientists explore and use complex neuroscience data.
The AQM Evaluation Suite is an automated framework based on RFC 7928 that simplifies simulation setup, topology creation, traffic generation, execution, and results visualization of AQM algorithms in ns-3. Originally developed in 2017 and maintained until 2019, the suite now needs an upgrade to: - Migrate its build system from waf to CMake - Integrate support for state-of-the-art AQM algorithms and enhanced ECN functionalities - Update simulation examples to reflect modern traffic control models and data center load balancing scenarios - Package and publish the suite on the ns-3 App Store
Xrefer currently functions as an IDA plugin for static binary feature extraction. The goal of this project is to abstract Xrefer’s core functionality into a flexible multi-backend pipeline supporting static analysis on IDA, Ghidra and Binary Ninja.
This project integrates Open Concept Lab (OCL) test concepts into OpenELIS-Global by parsing OCL ZIP files from a Configurations/ocl directory at startup, mapping them to the TEST table, and using checksums to process only modified files. It enhances lab standardization and interoperability, reducing manual setup time (e.g., 2 hours to 10 minutes) for technicians and ensuring consistent updates across sites for admins.
This project is about build and train a streaming chatbot on four datasets: Documentation site of cBioportal, Google group conversations, PMC papers used in studies, and OPENAPI. Also, the project used route logic to combine 4 chains, and wrote a routing logic function to choose chatbot depending on user query. The chatbot has chat history, indicators, and example questions, download button and footer (disclaimer) in user interface.
The objective of this project proposal is to enhance the performance of the existing patient flag module, update its code base to align with current opemrs standards, and introduce FHIR flag resource mapping. The current patient flag module has performance issue as it evaluates flags sequentially, this is leading to issues when dealing with multiple flags for a patient or group of patients. I propose evaluate these flag in parallel. This approach allows multiple flag evaluations to be performed concurrently, potentially improving performance compared to sequential evaluation. Furthermore, the module currently having methods with multiple database calls, such as saving multiple flags using a for loop, causing system slowdowns. To address this issue, I aim to reduce database calls through batch saving. The patient flag module does not compatibility with Java 11 and Java 17 and this module need to update to meet the current openmrs standards. Another objective of this project is to implement the FHIR flag resource in openMRS fhir module. In the project execution phase, first updating to Java 17 and revamping the patient flag module's code base, then addressing performance issues, and finally delivering FHIR flag resource mapping.
This project entails introducing support for measuring code coverage for all Kotlin files in the Android codebase with Bazel, with any support gaps documented.
The Robotics Academy is a framework designed for learning robotics and computer vision through a collection of robot programming exercises. Currently, the platform offers 26 exercises and 11 prototype exercises that students can access through the RoboticsAcademy docker image (RADI), which comes pre-installed with all necessary dependencies. The RADI is a key component of the platform, and this project aims to improve it by migrating it to Gazebo Harmonic from the current Gazebo11 version, which is nearing its end of life (2025). This migration will take form by having a couple of exercises running in ROS2 and Gazebo Harmonic with the new RADI.
In this project, we introduce quantum contrastive learning, a quantum machine learning technique to enhance performance of models involving high energy physics data for tasks such as signal and background classification. Contrastive learning is a way of producing effective representations of complex data such that it can be used for other downstream applications with minimal issues and increased throughput. High energy physics has several subtle aspects that usual images from other domains do not express, hence we conjecture that quantum computing might help getting improved representations for such data using entanglement and interference effects inherent to quantum mechanics.
Sustainability is a key goal for KDE over the next two years. As part of this goal, KDE aims to measure the energy consumption of its software and eco-certify applications with the Blue Angel ecolabel. Currently, we measure the energy consumption of our software using a lab based in Berlin following the process documented in the KDE Eco Handbook (Handbook: Applying The Blue Angel Criteria To Free Software). The lab setup requires (i) a test PC (System Under Test) to run user-emulation scripts, (ii) a power meter to measure the energy consumption of the test PC, and (iii) another PC (Data Aggregator & Evaluator) to collect and analyze the output data. In the current setup, developers must rely on individuals in Berlin to test software for them on-site in the lab. This proposal aims to provide remote access to the lab in two steps: first, automating the energy measurement process, including providing a summary of the results; and second, setting up backend CI/CD integration for measuring energy consumption. By automating the process and providing remote access, we make it possible for developers to measure the energy consumption of their software from any location in the world. The increased access to the lab will enable data-driven decision-making for efficiency improvements in code and software eco-certification with agencies like the Blue Angel. The proposed project has three main goals: i) Measurement Automation: Implement a solution where a developer provides as input three scripts (baseline, idle mode, and standard usage scenario) with a config file and automatically obtains a summary of results from OSCAR (Open source Software Consumption Analysis in R). ii) Remote Access 1: Implement a CI/CD pipeline as a backend solution for performing the energy consumption measurements in the remote energy measurement lab.
This project aims to automate the testing of stylus touch pens with Chromebook using a human motion robot (HMR). Currently, stylus pens or any other touchpad-related test cases take much longer time and are more expensive when going through lab testing. Developing and conducting these tests via a human motion robot will save half the cost and accelerate the testing time from 13 hours to 15 minutes. The objectives of this project include evaluating and assessing the pass/fail of stylus test cases, developing automation scripts for onboarding the HMR, and implementing the HMR’s testing lifecycle to run and evaluate test results. A three-step evaluation procedure has been proposed, and the onboarding automation scripts and the HMR lifecycle have been implemented. Evaluation results show successful evaluations with directions on future work aiming to improve and scale up the proposed approach.
The objective is to develop a machine learning model to tag sound effects in streams (like police sirens in a news-stream) of Red Hen’s data. A single stream of data can contain multiple sound effects, so the model should be able to label them from a group of known sound effects like a Multi-label classification problem. The first step would be to develop a baseline model using existing pre-trained deep learning models and add to the Red Hen’s pipeline. Then the performance can be improved using transfer learning and fine tuning the existing model to achieve better accuracy. In this process, the models can be trained on sound effects from noisy or human labeled data sets after they are pre-processed to avoid acoustic domain mismatch problems.
HPX being up to date with Std C++ Proposals, Senders/Receivers were implemented as per P2300. But they have been missing coroutine (co_await) integration and minor functionalities as described in P2300 which is likely to be accepted. Hence I propose to implement these functionalities within the Core HPX Library. Benefits: * Coroutines introduce better async code. For example, it is more readable, local variables have the same lifespan as the coroutine which means we don’t need to worry about allocation/release. * S/R algorithms can work with coroutines which they cannot as of now unless relied on futures which as mentioned are single-time use. * Adding co_await support makes the code more structured with respect to concurrency which can also be done by library abstractions of callbacks but using co_await may make it more optimized.
Merging duplicates and resolving conflicts is a major use case for JabRef users, according to a 2019 research conducted by JabRef developers. Currently, both features are implemented using a two-way merge UI. This often leads to a poor user experience as users cannot see the merged entry immediately. The goal of this project is to design and implement a new UI concept for merging entries that allow for a three-way merge. As well as implementing additional new features for making the merging process as easy and convenient as possible.
<p>Oppia's lessons can require between 15 minutes and an hour of time depending on a number of factors from learners: understanding of the material, literacy capabilities, and general focus. One potential behavior in the app is that we don't save the user's progress if they navigate away from an exploration (lesson) which, given how long lessons can take, is expected to be a frustrating experience. To mitigate this, we'd like to introduce support for a lightweight checkpointing system wherein users' saved state is fully retained if they navigate away from a lesson & back.</p>
<p><strong>Versatile Video Coding (VVC)</strong> is the latest video compression standard after HEVC. <strong>GStreamer need to support VVC, and we should start from a VVC parser</strong>. Codec parsers are a core part of the Gstreamer hardware decoder. They are used in <code>gstvappi</code> and <code>gstcodecs</code> to split frames, get slice level information. VVC specification is finalized on 6 July 2021. It's a good time to start work on the VVC codec parser for GStreamer.</p>
<p>Eclipse 4diac is an open source environment for programming distributed industrial automation solutions and control systems based on the IEC 61499 standard. One of the components of Eclipse 4diac is the 4diac IDE which is an integrated development environment for modeling distributed control applications compliant to the IEC 61499 standard. The current version of 4diac IDE does not support the validation of the models, which makes it difficult for users to get design-time feedback on inconsistencies regarding the models developed in 4diac. The Object Constraint Language (OCL) could be a solution to find issues in 4diac models since it provides capabilities for specifying generic constraints a model has to fulfill. The aim of this project is to develop OCL constraints and well-formedness rules to the metamodels of 4diac in order to improve the usability of the IDE.</p>
<p>App Inventor is a web application for building Android apps, which includes a block-based coding editor. The editor includes many blocks which accept enums as inputs. Currently these inputs are filled by assigning a constant number, which is not very readable nor localizable.</p> <p>This project seeks to replace those constants with enumeration dropdowns. These dropdown-blocks will be defined via <a href="https://en.m.wikipedia.org/wiki/Java_annotation" target="_blank">Java annotations</a>, so that they can also be added by <a href="http://ai2.appinventor.mit.edu/reference/other/extensions.html" target="_blank">App Inventor extensions</a>. They will then be added as <a href="https://developers.google.com/blockly/guides/configure/web/toolbox#shadow_blocks" target="_blank">shadow blocks</a> to the blocks in the <a href="https://developers.google.com/blockly/guides/configure/web/toolbox" target="_blank">toolbox</a>. In addition, other means of accessing the non-shadow versions of these blocks may be added.</p>
<p>After comparing various algorithms and brainstorming about their methods of implementation, I came up with the <strong>Rainbow</strong> <a href="https://arxiv.org/abs/1710.02298" target="_blank">Hessel et al., 2017</a> and <strong>Soft Actor-Critic</strong> <a href="https://arxiv.org/abs/1812.05905" target="_blank">Haarnoja et al., 2019</a> as the most in-demand and recent algorithms, whose implementation in mlpack would be crucial.</p> <p>Thus, I propose to add them to the existing RL codebase. Here are the details of what I expect to have accomplished at the end of the summer.</p> <ul> <li>Improving the current QLearning implementation. </li> <li>Implementing Rainbow as an improvement on DQN. This would include adding the following as extensions:<ul> <li>Dueling DQN</li> <li>Noisy DQN</li> <li>Categorical DQN</li> <li>N-step DQN</li> </ul> </li> <li>Writing test cases for each of the implementations </li> <li>Implementing Soft Actor-Critic (SAC) for continuous action space, along with its tests </li> <li>Creating detailed docs for all the above implementations </li> <li>Creating necessary environments, for proper testing of algorithms above (after discussing with mentors)</li> </ul>
<p>This proposal aims at three major things, improving the uBLAS interfaces that were added in GSOC 2019 work, stabilizing and enhancing the unit testing and C++20-ification of the library with new Concepts and better <code>constexpr</code> support.</p>
<p>Increasingly, distributed software development teams rely on online collaboration. The proposed project aims to implement the first skeleton of CoEditing in Che and Theia. In addition, the project would explore and address the synchronization challenges which arise in Online Collaboration. This could contribute to code conflicts reduction and to efficient management of common buffer while working with multiple language servers and editors.</p>
<p>OWASP Seraphimdroid has previously applied a system, based on permissions, which is able to distinguish malicious apps from non-malicious. But it still has some false positive errors, like foodpanda app. In order to improve OWASP Seraphimdroid’s performance, we would like to learn from other outputs (like network, CPU, buttery and memory usage, system call logs) can monitor about application whether it can be malicious.</p>