This repository contains the research code for the paper 'What's hidden in a randomly weighted neural network?' (arXiv 1911.13299). It implements methods for discovering high-performing subnetworks within randomly initialized neural networks without training the weights, building on the lottery ticket hypothesis. The codebase is organized around PyTorch model definitions (ResNet, custom ConvNets), config-driven experiment management (YAML), and modular trainers for different pruning strategies (default, lottery, sample-based sparsification). The core contribution is a set of custom convolution layers (SubnetConv, SampleSubnetConv, FixedSubnetConv) that learn binary masks over frozen weights via straight-through gradient estimators. The repository supports experiments on CIFAR-10, ImageNet, and MNIST, with extensive configuration files for small-scale and large-scale experiments. While the code is well-structured for research reproducibility, it relies on outdated dependencies (PyTorch 1.3.0, torchvision 0.4.1) and contains hardcoded paths, making it primarily useful as a reference implementation rather than a reusable library.
This repository is a curated bibliography and roadmap for Quantum Machine Learning (QML). It contains chronological lists of papers, books, and reviews from 2007 to 2026, organized into yearly markdown files. The content is entirely reference links with basic metadata. While there is no executable code, data pipelines, or analytical workflows, the repository serves as an authoritative, up-to-date resource for researchers entering or tracking the QML field. It also includes a workforce section and a separate reviews page. The repository is actively maintained with recent updates and follows a consistent structural pattern across years.
This repository is a tutorial and research exploration of Graph RAG (Retrieval-Augmented Generation) combining vector databases (Weaviate) and knowledge graphs (RDF/SPARQL). It contains Jupyter notebooks that demonstrate building knowledge graphs from PubMed articles, World Bank documents, and Netflix movies, comparing vector search and structured query performance for retrieval-augmented generation. A Streamlit app provides an interactive demo for medical article search with MeSH term refinement and LLM summarization. The code is primarily educational, with reusable query modules for Weaviate, RDF, and Databricks, but lacks dependency pinning and has hardcoded credentials, limiting reproducibility. The repository is hybrid, with both application (Streamlit app) and data research (notebooks, datasets) elements, but the dominant character is data research as the core value lies in the experimental notebooks and data pipelines.
This repository is a git mirror of the Emacswiki, containing the full version history of all wiki pages and uploaded files (primarily Emacs Lisp libraries) since August 2009. It is updated regularly via pushes from the Oddmuse wiki engine. There are no source code files or data pipelines; the entire content is a historical collection of wiki edits stored as text files in a git repository. The repository serves as an offline archive and version history for the Emacswiki community. While it is not a typical software or data science project, it is well-maintained and reliable for browsing or analyzing the wiki's history.
This repository functions as a curated "Awesome List" for Japanese Natural Language Processing (NLP), aggregating metadata from GitHub and Hugging Face. It includes a Claude Code plugin that allows searching the 1,197-resource dataset directly within the AI interface. The data lineage involves manual curation and automated scraping of external repositories, stored in `resources.json` and Markdown documentation. Critical risks include significant link rot (many "repo not found" entries), a suspicious future commit date (2026-07-20), and high technical drift due to the rapidly changing NLP landscape. While valuable for discovery, the static nature of the data and lack of environment definitions make it unsuitable for production reliance without verification.
This repository is a static Markdown document (README.md) that curates a list of companies known to use the Haskell programming language. It contains no executable code, data pipelines, or analytical models. The list is community-maintained via pull requests and includes company details such as location, sector, and remote work policy. As a reference resource, it is trivially reproducible by cloning the repository; however, it offers no reusable software components or data engineering artifacts. The data is presented in a structured table with cited sources per entry.
This repository contains the lab materials for the AWS Data Engineering Immersion Day, a guided workshop demonstrating how to build a data lake and analytics pipeline on AWS using services like AWS Database Migration Service (DMS), AWS Glue, Amazon Athena, Amazon QuickSight, Amazon Redshift Spectrum, and AWS Lake Formation. The repo includes step-by-step PDF lab guides, CloudFormation templates for infrastructure provisioning, sample data loading scripts, and a self-contained anomaly detection example using Apache Flink with Random Cut Forest. While the Flink streaming job is a real application, the primary purpose of the repo is instructional, not a production deployment or a self-contained research project. The data lineage involves ingesting ticketing data from a PostgreSQL database via DMS into S3, transforming with Glue into Parquet, and analyzing with Athena/Redshift. Clickstream data is generated into Kinesis streams and processed in real-time with Kinesis Analytics (Flink) to detect anomalies. The repo is hybrid: it has both application code (Flink Java) and data pipeline orchestration, but the dominant character is a workshop with infrastructure-as-code and documentation.
The FairFace repository serves as a documentation hub for the FairFace dataset, providing links to download facial images annotated with race, gender, and age attributes for bias measurement and mitigation in computer vision. It contains no source code, only a README.md detailing dataset characteristics, preprocessing methods (using dlib's get_face_chip with specific padding), and external links to download the data, labels, and pretrained models hosted on Google Drive. The repository functions purely as a reference point for accessing the dataset rather than offering any computational tools or reusable assets. As a data research artifact, it lacks essential components for reproducibility: no data loading scripts, preprocessing code, evaluation notebooks, or dependency management. Users must manually implement workflows to utilize the dataset. While the README provides clear attribution and download instructions, the absence of bundled data, environment specifications, or code limits its utility as a self-contained research resource. The dataset itself addresses bias in facial analysis systems, but this repository only facilitates access rather than enabling analysis. Maintenance appears stale based on the anomalous future-dated last commit (2026-01-04), indicating no recent activity. Technical drift is minimal due to the absence of code or dependencies, though external download links may risk obsolescence. The repository's value lies solely in its role as a citation pointer and access guide for the FairFace dataset, not in providing reusable data engineering components.
This repository hosts the GFWList, a community-maintained blocklist of domains and URLs blocked by the Great Firewall of China. The primary data asset is `list.txt`, a plain-text file in AutoProxy format containing thousands of rules (domains, IPs, patterns) used by proxy tools to bypass censorship. A GitHub Actions workflow automatically updates the checksum and date via a Perl script from the `apollyon` submodule, then base64-encodes the list into `gfwlist.txt` for distribution. The repository is purely a data source with minimal code; its value lies in the curated list itself, not in software engineering. It is widely used as a subscription feed for VPNs and proxy clients.
This repository implements an audio-driven facial animation pipeline that maps speech audio to blendshape weights for driving a digital human in Unreal Engine. The core is a deep learning model (Audio2Face) trained on paired audio (WAV) and blendshape weight data (exported from Maya). The training pipeline includes: (1) Linear Predictive Coding (LPC) feature extraction from audio via a C++ DLL, (2) data concatenation and selection of variable blendshape indices, (3) training a custom Keras model with a three-part loss (prediction, motion, emotion regularization), and (4) inference via TensorFlow Lite. A test application demonstrates real-time inference with microphone input, speech recognition (via AISpeech API), and UDP communication to UE4. The repository is primarily a research codebase with a demo application, but lacks dependency pinning, environment definitions, and contains hardcoded paths, making reproduction difficult. The model architecture and loss function are well-structured and potentially reusable.
This repository is a research artifact from the NL2Bash project (LREC 2018), providing a dataset and neural semantic parser for translating natural language descriptions into Bash one-liners. The core contributions include a parallel corpus of ~10k NL–Bash pairs, TensorFlow-based sequence-to-sequence models with attention and copy mechanisms, a custom Bash command parser (bashlint) that normalizes commands using a grammar extracted from man pages, and evaluation tools including tree edit distance and manual judgment pipelines. The codebase is primarily a research prototype with pre-trained model checkpoints, data preprocessing scripts, and evaluation files. While the dataset and parser are valuable, the dependency on TensorFlow 1.x, JavaCC, and Python 2/3 compatibility indicates significant drift. The repository is no longer actively maintained, but the documented methodology and reusable components (e.g., tree edit distance, statistical spell checker) remain relevant for similar NL-to-code tasks.
This repository provides two LiDAR degeneracy datasets (tunnel and fog) in ROS bag format, hosted on Hugging Face. It includes detailed metadata about the sensors (IMU, LiDAR, radar), their extrinsics, topics, and rates. The datasets are intended for research in LiDAR-radar-inertial fusion for challenging environments. The repository contains no code for processing the data, only documentation and links to the raw data.
This repository functions as a static dataset collection designed for prompt engineering and Large Language Model (LLM) evaluation. It contains no executable application logic, servers, or CLI tools. The primary data assets are structured JSON files (`prompts-zh.json`, `prompts-zh-TW.json`) and a fragmented corpus of over 600 Markdown files (`question/*.md`) containing tabular prompt data. The data lineage is direct: raw text prompts are curated into these files without ETL pipelines or transformation scripts visible in the repository. The repository serves as a benchmark and training corpus for Chinese-speaking AI users. It addresses the problem of prompt standardization and provides a wide range of task categories (classification, generation, translation, coding) for testing model capabilities. However, the data engineering structure is suboptimal; the use of 600+ individual Markdown files for a single dataset introduces unnecessary fragmentation and maintenance overhead compared to a consolidated CSV or JSON format. Maintenance activity is indicated by the provided commit timestamp (2026-04-28), suggesting the repository is actively updated. However, the content exhibits significant technical drift. The prompts heavily reference GPT-3.5 and specific time-sensitive data (stock prices, news events from 2022-2023), which reduces the long-term utility of the dataset without manual curation. The data is self-contained and requires no external dependencies to consume, making it highly reproducible for immediate use.
This repository is a static dataset of LeetCode coding problems curated by company and time period. It contains CSV files organized into directories named after hundreds of companies (e.g., Google, Amazon, Meta, Apple, Goldman Sachs), each with subfiles for different time ranges: 'Thirty Days', 'Three Months', 'Six Months', 'More Than Six Months', and 'All'. Each CSV row includes problem difficulty, title, a frequency metric (normalized so the top problem equals 100), acceptance rate, link, and topic tags. The data is a snapshot as of June 2025, derived from aggregated user reports or scraping. There is no executable code, no transformation scripts, and no automated update mechanism. Many CSV files are empty or contain incomplete data, indicating gaps in coverage. The repository serves as a reference for interview preparation but lacks reproducibility and documentation of data provenance.
This repository is a curated 'awesome list' for Federated Learning (FL) research, maintained as a Jekyll static site. It aggregates hundreds of linked GitHub repositories and DBLP search strategies. The core data asset is a machine-generated go.mod file that pseudo-pins the linked repositories as Go dependencies, produced by the add_dependent.py script which extracts GitHub URLs from README.md, queries the GitHub API for commit hashes, and writes a go.mod. However, this is not a functional Go module; it is a novelty lockfile. The dblp-search-keywords.md document provides a meticulous analysis of DBLP search syntax to discover FL papers without the 'federate' keyword. The repository has significant hybrid character: while it is primarily a content aggregation site (application-adjacent), its technical substance lies in the data exploration and search strategy documentation (research-oriented). The actual code is minimal and tightly coupled to the repository's specific structure, limiting reusability. The add_dependent.py script is a one-time scraper with hardcoded rate limiting and token placeholder, not a general-purpose tool.
This repository contains the legacy edition of 'Practical Numerical Methods with Python', a multi-campus MOOC (2014–2016) on numerical methods for science and engineering. It is organized as five stacked modules covering ODE integration (phugoid flight dynamics), finite-difference PDEs (convection, diffusion, waves), and elliptic problems (Laplace, Poisson, Stokes flow). The primary content is 26 Jupyter notebooks with interleaved theory, code, and assignments, plus helper Python modules and data files. The course is explicitly no longer updated; development has moved to a separate repository. The codebase is pedagogically strong but suffers from outdated dependencies (Python 3.4, NumPy 1.9.2, Numba 0.20.0), broken imports (numba.autojit removed), and reliance on a decommissioned Open edX platform for assessments. No tests, CI, or modern packaging exist. The repository is best treated as a historical archive and reference for numerical method patterns, not as a runnable course.
This repository is a compact PyTorch reimplementation of Siamese Networks for One-Shot Learning, trained and evaluated on the Omniglot dataset. It provides four core Python files: `model.py` defines a shared convolutional encoder with absolute difference scoring, `mydataset.py` handles in-memory image loading with rotation augmentations and positive/negative pair sampling, `train.py` orchestrates training and evaluation using gflags for configuration, and `make_dataset.py` flattens the Omniglot directory structure. The codebase uses PyTorch 1.0.1 and relies on deprecated APIs such as `torch.autograd.Variable`. Training achieves 89.5% accuracy on 20-way one-shot evaluation, slightly below the original paper's 92% due to optimizer choices (Adam vs SGD with momentum) and default parameter initialization. The repository also includes a static documentation site with bilingual (English/Chinese) support and a GitHub Pages deployment workflow. The research code is functional but tightly coupled to Omniglot and lacks modularity for reuse; the site is well-crafted but separate from the core ML logic.
The DSA-KIT repository is a static resource curation project containing no executable code or data processing pipelines. It functions as a comprehensive reference guide for Data Structures and Algorithms learning, organized entirely within a single README.md file that categorizes external links to books, courses, problem sets, visualization tools, and practice platforms by topic and difficulty level. The repository does not ingest, transform, or analyze any data; instead, it serves as a manually maintained directory of third-party educational resources. From a data engineering perspective, the project represents a flat-file dataset of resource metadata (titles, URLs, descriptions) structured using Markdown syntax. There is no ETL pipeline, automation, or validation logic—the curation relies entirely on manual updates by the maintainer. The architecture lacks separation of concerns, with all content (resource listings, categorization, and presentation) coupled in one file. While the resource organization follows a logical taxonomy (e.g., grouping by data structure type or learning objective), it demonstrates no novel data engineering patterns or reusable components. The repository shows signs of active maintenance with a recent commit (2026-03-18), but exhibits significant reproducibility risks due to its dependence on external links that may become inaccessible over time. There are no mechanisms for link validation, dependency tracking, or environment isolation. The core dataset (the list of links) is version-controlled and reproducible via cloning, but the practical utility decays as external resources change or disappear. Despite its popularity (537 stars, 72 forks), the project lacks the automation, scalability, and resilience expected of professional data engineering artifacts.
This repository serves as a technical deep-dive into Google's reCAPTCHA system, particularly focusing on its obfuscation techniques, fingerprinting mechanisms, and anti-tampering measures. It includes extracted JavaScript snippets that implement specific browser fingerprinting signals (e.g., storage access checks, window opener properties, WebGL data extraction) and JSON files containing example fingerprint data. The repository also contains analysis of the now-removed BotGuard component, including its obfuscated virtual machine implementations. While the subject matter is largely historical due to the removal of BotGuard in April 2026, the repository provides valuable insights into the techniques used by advanced anti-bot systems.
This repository is a curated list of resources for the React ecosystem, maintained as a README.md file. It contains links to various React-related projects, tutorials, libraries, and tools organized by category. There is no executable code or data processing logic; the value lies in the curated collection of links. The repository is popular and appears to be actively maintained based on the commit history.