Research & Education

NYU-DLSP21 is an open-source GitHub repository that provides comprehensive deep learning course materials from New York University's Spring 2021 semester. It includes lecture slides, Jupyter notebooks for practical exercises, and assignments, covering topics such as history, backpropagation, gradient descent, recurrent and convolutional networks, and latent variable energy-based models (LV-EBMs). The content has been reorganized and updated to reflect a more powerful and integrated approach to teaching deep learning, building upon LV-EBMs as a foundational module. This repository serves as a valuable resource for students and educators interested in advanced deep learning concepts, offering a structured curriculum and practical examples.

PINNs-Torch is a PyTorch-based implementation of Physics-Informed Neural Networks (PINNs), designed to accelerate scientific computing tasks. A key differentiator is its integration of CUDA Graphs and JIT Compilers (TorchScript), which can boost performance by up to nine times compared to earlier TensorFlow v1 implementations. The package is open-source and provides a robust framework for researchers and developers to build and experiment with PINNs. It includes examples for various problems, such as the Navier-Stokes PDE, and offers flexible installation options for both users and contributors. The tool is ideal for those looking to leverage the power of PyTorch for physics-informed machine learning, with a focus on speed and usability.

LongNet is an open-source implementation of the plug-in and play attention mechanism described in the paper "LongNet: Scaling Transformers to 1,000,000,000 Tokens." This Transformer variant is designed to significantly extend the sequence length that models can handle, reaching up to 1 billion tokens, while maintaining strong performance on shorter sequences. Its core innovation is dilated attention, which expands the attentive field exponentially as the distance between tokens grows. LongNet offers linear computational complexity and a logarithmic dependency between tokens, making it suitable for distributed training of extremely long sequences. Its dilated attention can be seamlessly integrated into existing Transformer-based optimization methods, providing a drop-in replacement for standard attention.

MatchSum offers an implementation of the ACL 2020 paper "Extractive Summarization as Text Matching." This tool is designed for researchers and developers working on natural language processing tasks, specifically extractive summarization. It supports both BERT and RoBERTa encoders and provides pre-trained models for the CNN/DailyMail dataset, as well as other datasets like WikiHow, PubMed, XSum, MultiNews, and Reddit. Users can process their own data by converting it to a specific JSONL format and generating candidate summaries. The code requires Python 3.7, PyTorch 1.4.0, fastNLP 0.5.0, pyrouge 0.1.3, rouge 1.0.0, and transformers 2.5.1, and is optimized for Linux environments with GPU support.

The MLOps Coding Course is an open-source educational resource designed to immerse learners in the practical applications of machine learning and artificial intelligence projects using Python. It caters to both beginners and experienced professionals, offering hands-on coding experiences and project-driven learning. The curriculum covers essential MLOps techniques, from environment setup and prototyping with Jupyter notebooks to productionizing code, ensuring quality through validation, and refining processes with CI/CD and containerization. The course also emphasizes sharing and observability for robust MLOps projects. It familiarizes users with industry-standard tools like uv, pytest, MLflow, and GitHub, and offers optional mentoring sessions and an MLOps Coding Assistant for enhanced learning.

minGPT offers a concise and educational PyTorch re-implementation of the OpenAI GPT (Generative Pretrained Transformer) model, covering both training and inference. Designed to be small, clean, and interpretable, it stands out from more sprawling GPT implementations. The core library consists of three files: `mingpt/model.py` for the Transformer model, `mingpt/bpe.py` for Byte Pair Encoding, and `mingpt/trainer.py` for PyTorch training boilerplate. It includes various projects and demos, such as training a GPT to add numbers or act as a character-level language model. While semi-archived, it serves as an excellent resource for understanding GPT's underlying mechanics before exploring more advanced versions like nanoGPT.

The Morocco IoT & AI Challenge is a regional capacity-building and pre-incubation program specifically designed for Senior University Students, High School Students, and Startups. It focuses on fostering innovative ideas within the fields of Internet of Things (IoT), Artificial Intelligence (AI), and related technologies. The program aims to stimulate and build the ecosystem supporting the Moroccan community by empowering participants through top-tier training, mentorship, media exposure, and networking opportunities. It offers an equity-free prize money and includes different tracks for high schools, graduation projects, and startups, guiding them through ideation, technical support, and demo days.

PokerGPT, rebranded as PokerOS, is an innovative AI poker coach designed to significantly improve a player's poker skills. Built by poker players for players, it offers personalized AI coaching that adapts to individual playing styles, helping users make better decisions at the table. The platform provides comprehensive training tools, ranging from fundamental concepts to advanced Game Theory Optimal (GTO) strategies. Users can elevate their game by mastering fundamentals, tracking progress with detailed analytics, and practicing specific real-world scenarios. It also includes GTO drills with a solver tool to boost skills using mathematical strategies. PokerOS is available on iOS, making it accessible for on-the-go learning.

Mock Exam AI is an artificial intelligence-driven platform designed to assist students and professionals in their exam preparation. The tool specializes in generating mock exams across a wide range of subjects and academic levels. Users can leverage its capabilities to create customized practice tests, providing a valuable resource for self-assessment and study. The platform aims to enhance learning outcomes by offering a structured way to simulate exam conditions and identify areas for improvement. It provides both free and premium options, catering to different user needs and offering a variety of features.

RTDL (Research on Tabular Deep Learning) is a comprehensive, open-source GitHub repository dedicated to advancing the field of deep learning for tabular data. It serves as a valuable resource for researchers and practitioners by curating a collection of academic papers and associated software packages. While the original `rtdl` Python package is deprecated, the repository itself remains active, pointing users to updated and more efficient packages like `rtdl_revisiting_models` and `rtdl_num_embeddings` for implementing models such as MLP, ResNet, and FT-Transformer. The project aims to provide up-to-date research and practical implementations, allowing users to stay informed on the latest advancements and apply deep learning techniques to tabular datasets.

ai-knowledge-graph is an open-source AI-powered tool designed to generate interactive knowledge graphs from unstructured text documents. It leverages large language models (LLMs) to extract Subject-Predicate-Object (SPO) triplets, identify entities, and infer relationships. Key features include automatic text chunking for processing large documents, LLM-assisted entity standardization to ensure consistent naming, and relationship inference to discover connections not explicitly stated. The tool supports any OpenAI-compatible API endpoint, offering flexibility with various LLMs like Ollama, LM Studio, and OpenAI. It generates an interactive HTML visualization with features like color-coded communities, node sizing by importance, and distinct representations for original versus inferred relationships, making complex information easily digestible.

Hustlr is an innovative iOS mobile application designed to help users hone their sales skills through AI-powered roleplay simulations. It provides a dynamic environment where individuals can practice their pitches against AI clients and receive immediate, actionable feedback and analytics to improve their performance. A unique aspect of Hustlr is its monetization model, allowing top-performing users to earn real money through Stripe, transforming practice into a competitive and rewarding experience. The platform encourages continuous improvement by enabling users to climb leaderboards and turn their enhanced skills into tangible financial rewards. It's ideal for anyone looking to refine their sales techniques in a practical, engaging, and financially incentivized setting.

wincnn is a Python module specifically designed to generate minimal Winograd convolution algorithms, which are crucial for optimizing convolutional neural networks. This tool implements the algorithms proposed in the research paper "Fast Algorithms for Convolutional Neural Networks" by Lavin and Gray (CVPR 2016). It provides symbolic computation capabilities, ensuring exact results for the transforms. Users can compute transforms for various F(m,r) configurations, including examples like F(2,3), F(4,3), and F(6,3), and also generate algorithms for linear convolution. The module requires Python 3.8 or higher and SymPy 1.9 or higher for its operation, making it a valuable resource for developers and researchers working on neural network optimization.

Bartleby is an online academic support platform designed to help students with their homework, textbook solutions, and writing needs. It offers access to thousands of textbook solutions written by subject matter experts and provides a platform for students to ask questions and receive step-by-step solutions from PhD and Masters-level experts. The tool also includes writing aids like paraphrasing and grammar checking tools to improve academic writing. Bartleby covers a wide range of subjects, including business, engineering, math, science, and social sciences, making it a comprehensive resource for various academic disciplines. Response times for expert Q&A can be as fast as 34 minutes for paid subscribers.

LingoSub is an AI-powered language learning platform designed to help users acquire new languages by watching videos with subtitles and AI-powered translations. It employs the comprehensible input method, exposing learners to content slightly above their current proficiency level but still understandable, promoting natural language acquisition. Users can tap on any word in the subtitles to see its definition and pronunciation, learning new vocabulary in context. The platform offers a vast library of videos, including content specifically curated for language learning and general entertainment, across 10 languages. LingoSub is accessible on desktops, tablets, and mobile devices, providing a seamless learning experience for beginners to advanced learners.

EasyDictation.app is an AI-powered language learning tool designed to help users master English listening and speaking skills. It transforms any YouTube video into an interactive lesson by generating AI transcripts with translations in over 70 languages. The app offers features like auto-pausing after each sentence, instant accuracy feedback with word-by-word comparison, and IPA pronunciation lookup. Users can also practice shadowing with AI pronunciation scoring, comparing their speech to native speakers. The tool automatically tracks difficult vocabulary, saving it to a personalized vocabulary book for review with flashcard games. Detailed progress reports, learning analytics, and a community leaderboard help keep learners motivated.

TrainrAI is an AI-powered fitness coach designed to personalize your workout journey by integrating directly with Apple Health. It analyzes key metrics from your Apple Watch, such as HRV, sleep patterns, and heart rate, to assess your daily recovery and training readiness. Each morning, TrainrAI delivers a clear, data-driven recommendation on the intensity of your workout, eliminating guesswork and optimizing your fitness progress. This intelligent system helps users set and achieve clear goals, providing actionable insights based on their unique health data. It's an ideal solution for individuals looking for a smart, personalized approach to fitness coaching that adapts to their body's daily state.

Analisi Logica Tool is an AI-powered online platform designed for the logical analysis of Italian sentences. It offers a fast, simple, and automatic way to break down phrases, providing examples, explanations, and a complete analysis. The tool is ideal for students, teachers, linguists, and language professionals, generating contextual analyses in real-time. Unlike traditional grammatical tools, it adapts dynamically to different sentence structures and is continuously updated based on linguistic research. Key features include voice command input, detailed breakdown of grammatical and logical elements like subject, predicate, and various complements, and instant, personalized data for each analyzed phrase.

Colorendo is an AI-powered coloring page generator designed to transform creative ideas into unique, printable coloring pages. Users can simply describe their idea, and the AI brings their vision to life in seconds. The platform caters to a wide range of needs, from playful animals and fairy tale scenes to simple patterns, making it suitable for all ages and skill levels. It offers features like organizing generated pages in chats, viewing creations in a gallery, and unlimited printing and downloading. Colorendo aims to inspire creativity and provide engaging activities for children and adults, making it ideal for parents seeking educational and fun activities.

Paird.ai is a collaborative AI tool designed to enhance pair programming experiences for development teams. It facilitates real-time code generation and allows users to create custom code suggestions and lessons. By selecting a specific programming language and setting a project goal, teams can leverage AI to streamline their coding process. The platform emphasizes seamless code collaboration and provides real-time feedback, making it easier for multiple developers to work together on coding projects efficiently. This tool aims to improve team productivity and code quality through intelligent assistance and shared development environments.

Awesome Full Stack Machine Learning Engineering Courses is a curated list of publicly accessible machine learning engineering courses from leading universities such as CalTech, Columbia, Berkeley, MIT, and Stanford. The resource also includes machine learning project case studies from large and experienced companies, with Python as the preferred language for end-to-end machine learning engineering. The list is organized by topics and areas of specialization, making it easy to navigate. It highlights highly recommended foundational courses and provides links to video lectures where available, covering subjects from computer science fundamentals to advanced deep learning, reinforcement learning, and natural language processing.

Amphion is an open-source toolkit designed for Audio, Music, and Speech Generation, aiming to support reproducible research and assist junior researchers and engineers in the field. It provides a unique feature: visualizations of classic models or architectures, which are beneficial for understanding complex models. The platform's objective is to offer a comprehensive solution for converting various inputs into audio, supporting individual generation tasks such as Text to Speech (TTS), Singing Voice Synthesis (SVS), Voice Conversion (VC), Accent Conversion (AC), Singing Voice Conversion (SVC), and Text to Audio (TTA). Additionally, Amphion includes several vocoders and evaluation metrics crucial for producing high-quality audio signals and ensuring consistent metrics in generation tasks. It also focuses on advancing audio generation in real-world applications, including building large-scale datasets for speech synthesis.

10 Academy is dedicated to preparing young African talent for global AI jobs through custom training programs and strategic partnerships. The platform collaborates with governments, universities, and companies to enhance employability and scale AI education. Key features include tenx, a scalable AI-enabled platform for personalized learning with real-time insights, and industry-aligned content in AI, data, and core career skills. The academy provides end-to-end support, from onboarding to job search, including tutor assistance, community engagement, and AI-powered job matching. 10 Academy aims to close the career-readiness gap and build the future of work in Africa, with a strong focus on economic impact and accessible training.

Explore Biology & Biochem Foundation Models is a Hugging Face Space dedicated to showcasing and exploring various biological and biochemical foundation models. While the live website currently indicates a runtime error, its intended purpose is to provide a platform for users to discover and interact with these advanced machine learning models. This tool is particularly valuable for researchers and educators in the fields of biology and biochemistry who wish to investigate and experiment with cutting-edge AI applications in their respective domains. It aims to facilitate understanding and application of complex biological data through the lens of foundation models.