Explore Projects

<p>Nuitka is a Python compiler written in Python. It is a seamless replacement or extension to the Python interpreter and compiles every construct that CPython does. Nuitka works by translating Python code into a C level program which can be executed in the same way as CPython using libpython and a few C files of its own. All optimizations of Nuitka are aimed at improving performance while ensuring perfect compatibility. This project ensures Nuitka’s compatibility with the top 50 PyPI packages by setting up automated testing for each package. These automated tests will serve to be very important tools for the development of Nuitka, as they will be used to assure Nuitka’s compatibility with the most used Python packages every time Nuitka receives a new update.</p>

The proposal consists in treating the time series creation time as another metadata. We propose a few incremental improvements that can be released separately, and therefore faster. We start by dropping time series when negotiating OpenMetrics format. Then changing the current metadata behavior to be correlated with MetricPoint instead of MetricFamily. The third step is storing the last seen value of _created for each MetricPoint. Lastly, storing a symbolic value each time the _created timestamp minus the scrape timestamp is lower than the scrape interval.

<p>This project aims at firstly, investigating the performance of the current implementations against its alternatives at various levels of data transfer over the network connection for Duet, and further integrating the improvements into the PySyft stack.</p>

ScummVM includes a global fully configurable keymapper, but this requires engines to be adapted to use it. Hence the goal of this project is to integrate the customised ScummVM keymapper into the engine's input handling system. This involves modifying the engine-driven input handling code for mapping user input to in-game actions.

Lately various AI driven algorithms have been published that are able to extract single instrument recordings (STEM) from an audio track file. Native Instruments has introduced an open definition for stem files STEMS to store these extracts. A Stem file contains the original a stereo track and four additional channel with the musical elements: A drums stem, a bassline stem, a melody stem and a vocal stem, in a m4a container. Currently Mixxx is only capable of playing the stereo track. The goal of this project is to extract all channels and allow the user to mix them individually. In a fist version the Stem mixer can replace the equalizer in the Mixxx GUI. In an advanced version an integration with effects system is desired.

<p>The idea is to organise each of the solvers into a class with a consistent interface. Then the main part of dsolve can loop over these classes rather than having a monolothic matching function.</p> <p>This will make the code easier to maintain and understand and eliminate the need for global variables.</p> <p>This can also make dsolve much faster. Currently all the matching code for all of the methods runs for every ODE before a solution is returned. Some of the methods can return a solution much faster than other methods can run their matching code. Once everything is neatly organised into classes the main match and solve code can be changed to return the first solution that matches.</p>

The PDK or (plugin development kit) aims to allow loading artifacts into tremor at runtime instead of requiring a full recompilation. This streamlines the development process of extending tremor with custom functionality. The goal is to allow live loading of a number of artifacts so start with: Pipeline Operators, Codecs, Pre- and Postprocessors, Custom Functions.

<p>Zero-knowledge proofs have an important role to play in the future of verified machine learning prediction. However, no deep learning framework has the ability to perform verified computation of neural networks using ZKPs. In this project, I will wrap an existing ZKP library into a new tensor type allowing users of PySyft to generate, evaluate, and verify tensor computations, Here I will port an existing ZKP library, libsnark into PyTorch operations because there is no doubt that C++ is faster than Python and can perform a huge number-crunching job so fast and efficiently.</p> <p>This library implements Non-Interactive, zkSNARK schemes, which are a cryptographic method for proving/verifying, in zero-knowledge, the integrity of computations. A computation can be expressed as an NP statement, A prover who knows the witness for the NP statement (i.e., a satisfying input/assignment) can produce a short proof attesting to the truth of the NP statement.</p>

I believe Qubes is an easy and intuitive tool to achieve “reasonable” security, as it relies on explicit compartmentalization and force users to actively think about their actions. As massive cyberattacks are growing, Qubes is increasingly becoming an essential tool for the general public, reinforcing the need for UX- focused software. This project will make Qubes more accessible for people from diverse backgrounds by providing UX-focused GUI tools for software installation in templates.

<h6>This project aims at improving the flexibility and resource management of Apache Nemo by dispatching Nemo executor to AWS Lambda Function.</h6>

<p>Command line is the most flexible and powerful platform. It is also the fastest way to work compared to a GUI. Also, the command line will give scope for automation and scripting. Currently, the command line handler framework in Pharo is used. Pharo comes with many built-in handlers. ​ There is also Scale​ repository which is also used to write scripts. Due to the increased use of scale, the complexities in Pharo has been raised to a higher level leading to core problems in Pharo. I wish to solve these problems, thereby improving the command line management.</p>

Network topologies and graph databases go hand-in-hand. The OpenAPI specifications for Kubernetes provides taxonomy, but augmenting a graph data model with formalized ontologies enables any number of capabilities, one of the more straightforward is the inferencing requisite for natural language processing, and consequently, a human-centric query / response interaction becomes becomes possible. More importantly, more advanced systems can be built when a graph data model of connected systems is upgraded to be a knowledge semantic graph. Deliverables (among other items): - a Kubernetes ontology using OWL as a popular (and mature) way of doing this. - a cuelang-based component generator

<p>This project aims to implement an iterator over k- shortest source to destination paths using improved versions of Yen’s algorithm for simple paths. Apart from working on these methods the scope of this project will be improving and cleaning the methods that are currently too slow in python as these methods will be a dependency for implementing the other algorithms.</p>

<p>FV (Fan-Vercauteren) Homomorphic Encryption scheme is one of the leading approaches in homomorphic encryption. Homomorphic encryption is a form of encryption that allows computation/operations on ciphertext(encrypted data), generating an encrypted result which, when decrypted, matches the result of the operations as if it had been performed on the unencrypted data. This can allow critical data to be used by the third party even without letting them know what actually data was and ensuring data privacy.</p> <p>So, My proposal is to implement the FV homomorphic encryption scheme inside PySyft library because the algorithms which are natively implemented in PySyft can automatically run across Python, GPUs, Javascript, Android (Kotlin), and iOS (Swift). In order to meet up the requirements, and realize my proposal, I propose a tensor of type FVTensor which will have the ability to perform the FV homomorphic encryption operations elegantly.</p>

<p>Over the years, <code>Deep Learning</code> has become a promising field of work, attracting attention from the most prominent <code>Machine Learning</code> researchers of the world. One of the most prominent ideas in the field of <code>Deep Learning</code> is <code>Generative Adversarial Networks</code> invented by <code>Ian Goodfellow</code>. This proposal aims to implement <code>Deep Convolutional GAN (DCGAN)</code> and <code>Wasserstein GAN (WGAN)</code>. In addition to these, some additional work is also planned to be done, namely the implementation of <code>Stacked GAN (StackGAN)</code> and <code>Deep Belief Networks (DBN)</code>. The rest of the details are mentioned in the proposal.</p>

During GSoC'24, a huge effort was made to improve Checkstyle's Google config coverage and updated our implementation of google style guide to the latest version ( 03 Feb 2022 ). but there's still a lot of work left. google_checks.xml is the configuration file where our google java style guide is implemented. Though most of the implementation work has been done and we have covered almost all the rules in google java style guide, users have reported a bunch of issues pointing out flaws in our implementation, these issues are labeled as google style issues, we need to solve them. On top of this, we have few rules which we are not able to completely follow/implement, they are marked with "Blue Tick" in our coverage page, we need to find a way to reduce such blue ticks rules and improve our coverage

This project aims to update the current memory allocator used by RTEMS to be modular, allowing users to provide their own heap implementation or use one supplied by RTEMS. It will also include implementing the TLSF allocator as the default allocator instead of the current first-fit approach. This change will address the existing fragmentation and unpredictable runtime issues associated with the current allocator. Additionally, the project will explore the implementation of RT-Mimalloc, a modified version of Microsoft’s mimalloc optimized for real-time systems, representing the first open-source realization of this allocator. Finally, the project will add an allocator benchmark to evaluate the new memory allocators as well as any user-provided implementations.

<p>The Webpack Dev Server is designed to make development of Webpack projects efficient, customizable, and possible on older browsers. During this project, I will introduce native WebSockets on the client and ws on the server as the new default Dev Server implementation, while also allowing the user to provide their own client-server implementation. I will improve upon API/CLI uniformity, which has made the Dev Server difficult for users to work with in the past. Finally, I will fully implement the Dev Server into webpack-cli, and I will create core Webpack features that allow the Dev Server to more easily interact with the Webpack Compiler.</p>

Latency Tolerance Reporting is used to help manage ASPM (Active State Power Management). Some of the LTR values are platform-dependent and must be learned from platform firmware. Linux currently does not ask firmware for this information. I don't know exactly how this information needs to be incorporated into the ASPM support, but I think it's something we should figure out. See the PCI Firmware Spec, r3.2, sec 4.6.6.

A collaboration server is owned by Mission Support System. Local users can be created by using this server. Existing identity providers using SAML 2.0 are desired to be used. A service provider (SP) needs to be implemented on the server side in the existing WSGI application and an authentication into the QT client application. When a user logs in on the QT client, a browser will be triggered for the login process. The QT client user can authenticate afterward by exchanging a one-time token. To test the system, a local identity provider (IdP) will be configured and a few tests will be added.

WasmEdge supports PyTorch, TensorFlow Lite, llama.cpp, and more NN backend. Dealing with the text-to-voice is a big thing that we want to achieve. To make it possible, we would like to integrate piper (https://github.com/rhasspy/piper), A fast, local neural text-to-speech system in C++ as a new WASI-NN backend (https://github.com/second-state/wasmedge-wasi-nn). Expected Outcome: 1. A new plugin provides a piper WASI-NN backend 2. A test suite for validating the plugin 3. Documents and examples for explaining how to use the plugin.

This project utilizes Hyperledger Fabric to develop a federated learning framework with committee consensus. Hyperledger Fabric is used for storing and tracing global model update exchange. A customized committee consensus mechanism is required to reduce a load of consensus computation by selecting a random smaller subset of nodes to participate in consensus each round. In order to protect the identity of local data owners, the proposed framework will be integrated with Indy, Aries, and Ursa stack projects to offer self-sovereign identity. The mentee will work with the project mentors to gather and validate the requirements, and design the appropriate technical solution.

<p>This new package aims to fit into the tidyverse and grammar of graphics suite to support and facilitate temporal-context data analysis and visualisation. A new class tbl_ts will be defined and provided for representing regular time series in a tidy data structure. Some common verbs from the dplyr package, such as filter, summarise, and mutate, will be inherited to transform temporal data more easily and intuitively. New calendar-based arrangements created using linear algebra tools will be developed to display temporal data in a format that captures people’s daily schedules. A new faceting system that attempts to reflect a nested structure will be experimented, which would provide a visualisation framework for hierarchical and grouped time series data.</p>