Research & Education

what_are_embeddings is an open-source GitHub repository dedicated to exploring the fundamentals, history, and industrial usage patterns of embeddings in machine learning. The project includes a comprehensive LaTeX document, a generated website, and supporting notebook code, making it a valuable resource for anyone looking to understand these numerical representations. It covers the evolution from traditional methods like TF-IDF and PCA to modern approaches enabled by Word2Vec and Transformer architectures. The repository is designed for educational purposes, offering a deep dive into how embeddings scale with increasing data volume, velocity, and variety in modern applications. Users can contribute to the document by building the LaTeX artifact and submitting pull requests.

BookAI is a mobile application developed by Big Mobile Games that leverages artificial intelligence to transform user ideas into professional, personalized e-books. With its intuitive interface, users can generate complete stories and books with just a few clicks. The application is designed to be user-centric, delivering simple, original, and useful experiences. Launched in September 2023, BookAI is available for free download on Google Play, making it accessible for anyone looking to quickly create unique content for personal use or sharing. It aims to simplify the book writing process, allowing users to take full author credit for their AI-generated creations.

WebGLM is an efficient web-enhanced question-answering system developed by THUDM, presented at KDD 2023. It leverages a 10-billion-parameter General Language Model (GLM) to integrate web search and retrieval capabilities, significantly improving the accuracy and relevance of answers. The system features an LLM-augmented Retriever for enhanced web content retrieval, a Bootstrapped Generator for human-like response generation, and a Human Preference-aware Scorer to ensure useful and engaging content. WebGLM supports both 2B and 10B parameter models and offers options for searching via SerpAPI or Bing. It is designed for researchers and developers looking to implement advanced, web-aware QA systems.

Mathematics-for-ML is a comprehensive collection of resources designed to help individuals learn and review the mathematical foundations necessary for machine learning. This GitHub repository offers a curated list of books, academic papers, and video lectures covering key areas such as algebra, topology, differential calculus, optimization theory, linear algebra, probability, and statistics. It serves as an excellent starting point for anyone looking to build a strong mathematical understanding for deep learning and machine learning concepts. The resources range from introductory texts to more advanced topics, making it suitable for various learning levels.

vqa.pytorch is an open-source project offering a PyTorch implementation for Visual Question Answering (VQA). Developed by researchers at LIP6 and Heuritech, this tool aims to facilitate the reproduction of state-of-the-art results, particularly those achieved with the MUTAN: Multimodal Tucker Fusion for VQA method on the VQA 1.0 dataset. It provides a modular and efficient codebase for further research on various VQA datasets. Key features include support for different VQA datasets (VQA 1.0, VQA 2.0, VisualGenome), pretrained models, and tools for extracting features from images using convolutional neural networks. The repository also includes documentation on its architecture, options, and quick examples for training and evaluating models, making it a valuable resource for researchers and students in the field of computer vision and natural language processing.

The MATLAB-Simulink-Challenge-Project-Hub offers a curated collection of research and design project ideas, enabling users to gain practical experience and insight into current technology trends and industry directions. These projects are designed to contribute to engineering and science by addressing key industry challenges. Participants can explore various topics like Artificial Intelligence, Autonomous Vehicles, Robotics, and Sustainability. Upon project completion, individuals can receive official recognition from MathWorks research leads and rewards. The hub also provides an interactive filtering tool to help users find projects matching their interests by technology, application area, and MATLAB/Simulink tools. It emphasizes open and accessible work, with guidelines for generative AI use.

AI Denmark is a dedicated initiative aimed at empowering Danish small and medium-sized enterprises (SMEs) to effectively integrate artificial intelligence and data-driven strategies into their business operations. The platform provides comprehensive support, including insights through articles, videos, podcasts, and case studies, as well as community events and a newsletter. Businesses can apply for a tailored program that offers customized assistance to explore and implement AI solutions. The vision is to position Denmark as a leader in the responsible and business-oriented application of AI, fostering competitive advantages and future-proof business models for its SMEs.

The Clay Foundation Model is an open-source AI model and interface specifically designed for Earth observation and environmental modeling. It offers a comprehensive toolkit for developers and researchers to work with Earth-related data, including digital elevation models, Sentinel-1, and Sentinel-2 satellite imagery. The model and its source code are licensed under the Apache-2.0 license, ensuring broad usability and modification, while its documentation is licensed under CC-BY-4.0. The project provides clear installation instructions via pip or mamba, and detailed guidance on running and training the neural network model using LightningCLI v2. It supports development and deployment in JupyterLab environments, making it accessible for scientific computing tasks.

GrokiPediaVerified is an open-source knowledge base designed to be a comprehensive collection of information across various subjects. It allows users to actively contribute by suggesting new articles or proposing edits to existing content. The platform encourages specific and well-sourced contributions, emphasizing quality over quantity. Users can sign in to manage their suggestions and edits, fostering a community-driven approach to knowledge curation. It aims to centralize information for efficient exploration, covering topics from artificial intelligence to historical figures and scientific breakthroughs.

Citrus is a similarity-based search engine designed for scientific literature, leveraging machine learning to identify and present related papers. It offers researchers a comprehensive overview of relevant articles within a specific research field through a single search query. This tool aims to streamline the literature review process by helping users quickly discover interconnected research, making it easier to identify key studies and trends. By focusing on similarity, Citrus assists in navigating vast academic databases efficiently, ensuring researchers can find pertinent information without extensive manual sifting.

ML-foundations is a comprehensive educational resource, developed by Jon Krohn, that provides foundational knowledge essential for understanding contemporary machine learning approaches, including deep learning and other AI techniques. The curriculum is structured into eight subjects across four core areas: Linear Algebra, Calculus, Probability and Statistics, and Computer Science (Algorithms & Data Structures, Optimization). It offers practical, functional understanding through vivid illustrations, paper-and-pencil exercises with solutions, and hundreds of Python code examples in hands-on Jupyter notebooks, primarily focusing on PyTorch and TensorFlow libraries. The content is accessible through various channels including YouTube, O'Reilly, Udemy, and ODSC, catering to different learning preferences and offering additional features like interactive testing and certificates through paid options. It is ideal for data scientists, software developers, and AI enthusiasts looking to deepen their understanding of ML fundamentals.

ML-paper-notes is a comprehensive GitHub repository dedicated to providing concise notes and summaries of significant research papers across machine learning, computer vision, and natural language processing. Organized by subject, the repository offers PDF summaries for each paper, making it an invaluable resource for researchers, students, and practitioners looking to quickly understand complex topics without sifting through entire papers. The collection covers a wide range of areas including Self-Supervised & Contrastive Learning, Semi-Supervised Learning, Video Understanding, Domain Adaptation, Explainability, NLP, Generative Modeling, Semantic Segmentation, and more. Each entry links directly to the original paper and its corresponding notes, facilitating efficient academic review and knowledge acquisition.

ml-powered-applications serves as the official companion code repository for the O'Reilly book 'Building Machine Learning Powered Applications'. This open-source project offers a practical resource for software developers and machine learning engineers looking to understand and implement ML-powered applications. The repository includes a comprehensive set of Jupyter notebooks that illustrate key concepts covered in the book, a Python library (`ml_editor`) containing core functions for a Machine Learning-driven writing assistant case study, and a Flask application demonstrating how to serve ML results to users. It supports Python 3.6 and 3.7, with setup instructions for virtual environments and necessary package downloads (spaCy, NLTK).

CompSciLib is an AI-powered study tool specifically developed to assist students enrolled in mathematics and computer science programs. The platform's primary goal is to significantly reduce study time, potentially by half, while simultaneously contributing to improved academic performance. It offers a suite of AI-driven functionalities tailored to support various aspects of learning within these technical fields. While specific features are not detailed on the current website, the tool's focus is on providing comprehensive AI assistance for complex subjects, suggesting it likely includes capabilities such as problem-solving assistance, concept explanation, and potentially practice material generation to enhance understanding and retention.

NABLA-SciML is a comprehensive collection of ongoing work in Scientific Machine Learning (SciML), developed by Juan Diego Toscano under the mentorship of Prof. George Karniadakis. This repository serves as a unified framework for efficient and reproducible implementations of various Physics-Informed Neural Networks (PINNs), DeepONets, and newer architectures like KANs. It includes tutorials for both PyTorch and JAX, along with specific modules for Residual-Based Attention (RBA), comprehensive KANs (cKANs), Kurkova-Kolmogorov-Arnold Networks (KKANs), and a Variational Framework for Residual-Based Adaptivity (vRBA). The project also features a custom, highly accurate implementation of the Self-Scaling Broyden (SSBroyden) optimizer, making it a valuable resource for researchers and students in the field.

Narrative Nooks leverages AI to deliver engaging and personalized learning experiences through interactive stories. Designed for young learners, the platform offers a wide variety of subjects and features to help children master new skills. Key offerings include dynamically generated lesson badges, custom story creation, and on-demand audio and image generation. Users also benefit from 24/7 on-call tutoring support, ensuring expert help is always available. The platform aims to captivate and educate through narrative-driven lessons, making learning both effective and enjoyable for students globally.

Toritark is an AI-powered language learning platform designed to help users improve their language skills through personalized, level-appropriate stories. It supports 18 languages and offers features like sentence-level translation and comprehension assistance to facilitate understanding. Users can actively engage with the content by retelling stories and receiving AI feedback on their grammar, vocabulary, and style, which is crucial for practical application and improvement. The platform also incorporates vocabulary management and spaced repetition techniques to strengthen long-term memory and enhance speaking fluency, making it a comprehensive tool for language acquisition.

MOSS-RLHF is an open-source project from OpenLMLab that delves into the intricacies of Reinforcement Learning from Human Feedback (RLHF) within large language models, specifically focusing on the Proximal Policy Optimization (PPO) algorithm. The project received the best paper award at the NIPS 2023 Workshop on Instruction Tuning and Instruction Following. It provides researchers with competitive Chinese and English reward models, which boast good cross-model generalization abilities, reducing the need for extensive human preference data relabeling. The project also offers in-depth analysis of the PPO algorithm, proposing the PPO-max algorithm for stable model training, and releases complete PPO-max codes to help align LLMs with human preferences. It includes resources for training reward models and policy models, along with annotated datasets.

YouTube Summarize is an AI tool hosted on Hugging Face designed to provide concise summaries of YouTube videos. This tool allows users to quickly extract the main points and key information from video content, eliminating the need to watch lengthy videos in their entirety. It is particularly useful for students, researchers, and professionals who need to efficiently process information from video lectures, tutorials, or presentations. The tool is free to use and operates as a Hugging Face Space, offering a straightforward way to get video summaries without complex setups or subscriptions. While the specific features beyond basic summarization are not detailed, its core function is to streamline information consumption from YouTube.

node2vec provides a Python3 implementation of the node2vec algorithm, designed for scalable feature learning in networks. It allows users to generate node embeddings from graphs, which can then be used for tasks like node classification, link prediction, and visualization. The tool supports various parameters for customizing the embedding process, including dimensions, walk length, and the number of walks per node. It also offers functionality for embedding edges using methods like Hadamard, Average, WeightedL1, and WeightedL2. The implementation is open-source and integrates with `gensim.Word2Vec` for model fitting and vector operations, making it a powerful tool for researchers and practitioners working with graph data.

Progerente is a corporate software development company that leverages technology to empower businesses. They specialize in developing innovative solutions across several key areas, including artificial intelligence, augmented and virtual reality, and web and app platforms. Their services are designed to meet the strategic needs of various organizations, from enterprises and universities to banks, hospitals, and government entities. Progerente's approach involves a strategic vision and methodological development, ensuring impactful digital solutions. They have a proven track record, with leading companies in Latin America as clients, and have been featured in numerous media outlets for their innovations in business empowerment.

nndl offers a Chinese translation of Michael Nielsen's renowned online book, "Neural Networks and Deep Learning." This project serves as a valuable educational resource for Chinese-speaking individuals interested in the fundamentals of neural networks and deep learning. The translation is meticulously crafted in LaTeX, ensuring a high-quality visual presentation of mathematical equations and plots, which is crucial for understanding complex technical concepts. It also incorporates finished work from another translation project, demonstrating a collaborative approach to providing comprehensive learning materials. Users can compile the LaTeX source code to generate a PDF document, requiring a TeX system and specific fonts for optimal display.

neuralhydrology is an Open Source Python library designed for training neural networks with a strong emphasis on hydrological applications. This package has been extensively used in research and various academic publications, highlighting its utility in the field. The core principle of the library is modularity, allowing for easy integration of new datasets, model architectures, and training-related aspects such such as loss functions, optimizers, and regularization techniques. It is built on the deep learning framework PyTorch, known for its flexibility in research. The library supports configuration files, enabling users to train neural networks without directly modifying the code. It is actively maintained by the AI for Earth Science group at the Institute for Machine Learning, Johannes Kepler University, Linz, Austria.

Godela is an AI Physics Engine designed to revolutionize engineering and research by applying AI to understand the physical world. It learns the governing behavior of systems from existing simulation data, allowing users to train physics-constrained models. With Godela, users can explore 'what-if' scenarios by changing parameters like airflow, geometry, or materials, and receive physics-accurate answers in seconds, significantly faster than traditional simulations. The tool then helps converge on optimal configurations, turning months of R&D into minutes. Godela's applications span various physical domains, including data center thermal optimization, electronics cooling, aerodynamics, electromagnetics, and structural analysis, enabling faster problem-solving and design iteration.