Research & Education

Goopt is an experimental search engine designed to simulate a "procedural web" using GPT-3. This tool explores the concept of Web 4.0, where content is generated algorithmically and synthetically by artificial intelligence. Users can search for any term, receive related results, and access generated content, mimicking the experience of browsing a traditional web. The project highlights the potential for infinite, personalized content generation through natural language processing and procedural content generation, adapting to user queries and preferences. Currently, Goopt generates text content and requires users to replicate the project locally with their own OpenAI API Key, as there is no public online version.

World Pulse Now is an AI-powered platform designed to keep users informed about the latest global trends, news, and insights in real-time. It aggregates news from various sources, presenting it on a single, powerful page for quick consumption. A key feature is its global trends heatmap, which offers an insightful visual overview of the world’s most important stories at a glance. The platform categorizes news into areas like U.S. News, World Affairs, Financial Markets, Cryptocurrency, Technology, Artificial Intelligence, Science, Health, Sports, and Entertainment. It also provides options to filter stories by sentiment (Positive, Neutral, Negative) and language (English, Français, Español, العربية). Additionally, WP Now offers a "WPN Daily Briefing" podcast for quick daily stories, making it easier for users to stay updated without information overload.

ResShift is an efficient open-source diffusion model designed for image super-resolution, developed by Zongsheng Yue and others. It addresses the common limitation of slow inference speeds in diffusion-based SR methods by introducing a novel residual shifting technique, which drastically reduces the required sampling steps to as few as 15, or even 4 in its journal version, without compromising output quality. This approach constructs a Markov chain that efficiently transfers between high-resolution and low-resolution images. Beyond super-resolution, ResShift also supports applications like image deblurring, natural and face image inpainting, and blind face restoration. The project has been recognized at NeurIPS 2023 (Spotlight) and published in TPAMI@2024, highlighting its advanced capabilities and efficiency in image enhancement.

Roadmap-To-Learn-Agentic-AI is an open-source GitHub repository offering a comprehensive guide to mastering agentic AI systems. It begins with foundational knowledge in Python programming and essential machine learning concepts, including Natural Language Processing (NLP) techniques like TFIDF and Word2vec. The roadmap then progresses to in-depth Deep Learning for NLP, transformer explanations, and extensive Generative AI tutorials with end-to-end projects. A significant portion is dedicated to Agentic AI tutorials, exploring various frameworks such as Langchain, LangGraph, Agno, Phidata, CrewAI, and Autogen. This resource is ideal for individuals looking to build a strong understanding and practical skills in the rapidly evolving field of agentic AI.

AI-PRO.org serves as a comprehensive resource for individuals looking to navigate and understand the rapidly evolving world of artificial intelligence. The platform offers a curated collection of AI applications and educational materials designed to help users grasp current AI trends, technologies, and best practices for utilizing generative AI. By joining AI-PRO, users can explore various AI tools that enhance creativity and productivity. The platform aims to provide clarity and confidence in using AI, making it an essential hub for anyone interested in learning about or applying AI in their daily tasks or professional endeavors. It emphasizes practical application and understanding of AI's power.

Read Frog is an open-source, AI-powered browser extension designed to transform everyday web reading into an immersive language learning journey. It supports seamless switching between bilingual and translation-only modes, providing context-aware AI translation by extracting page titles and content summaries for enhanced accuracy. Users can select any text for instant translation, detailed explanations, or text-to-speech playback. The tool allows for custom translation prompts using dynamic tokens and optimizes API costs with intelligent batch requests. It connects to over 20 AI providers, including OpenAI, Claude, and Google Gemini, and offers free translation options like Google Translate. Additionally, Read Frog provides subtitle translation for YouTube videos and high-quality text-to-speech with 150+ voices across 80+ languages, making it a comprehensive solution for language learners.

RoleLLM-public is a comprehensive framework designed to benchmark, elicit, and enhance the role-playing capabilities of Large Language Models (LLMs). It introduces RoleLLM, a four-stage process encompassing role profile construction, Context-Based Instruction Generation (Context-Instruct) for knowledge extraction, Role Prompting using GPT (RoleGPT) for style imitation, and Role-Conditioned Instruction Tuning (RoCIT) for fine-tuning open-source models. The framework includes RoleBench, a systematic and fine-grained character-level benchmark dataset with over 168,000 samples. RoCIT on RoleBench has led to the development of RoleLLaMA (English) and RoleGLM (Chinese), significantly improving role-playing performance to levels comparable with GPT-4.

Hivemind is a PyTorch library designed for decentralized deep learning, facilitating the training of large models across a vast network of computers, including those from different universities, companies, and volunteers. Its core features include distributed training without a master node, leveraging a Distributed Hash Table for network connectivity. The library ensures fault-tolerant backpropagation, allowing training to continue even if some nodes become unresponsive. It also implements decentralized parameter averaging, enabling iterative aggregation of updates from multiple workers without requiring network-wide synchronization. Furthermore, Hivemind supports training neural networks of arbitrary size by distributing parts of their layers using Decentralized Mixture-of-Experts. It is compatible with Python 3.8+ and PyTorch 1.9.0+ and offers integration with PyTorch Lightning for adapting existing pipelines to slow, unreliable networks.

Hands-on-transfer-learning-with-python is a comprehensive GitHub repository designed to simplify deep learning through the application of transfer learning techniques. It leverages the Python deep learning ecosystem, including TensorFlow and Keras, to provide practical examples and code. The resource is structured into three main sections: deep learning foundations, essentials of transfer learning, and transfer learning case studies. It covers important deep learning architectures like CNNs, DNNs, RNNs, LSTMs, and capsule networks, and explores state-of-the-art pretrained networks such as VGG, Inception, and ResNet. The repository includes real-world case studies in computer vision, audio analysis, and natural language processing (NLP), making it an invaluable resource for practitioners looking to implement advanced deep learning models.

spaCy is a powerful, open-source library for advanced Natural Language Processing (NLP) in Python and Cython. Designed for production use, it incorporates the latest research and provides pre-trained pipelines for over 70 languages, enabling tokenization and training. Key features include state-of-the-art speed, neural network models for tasks like tagging, parsing, named entity recognition, and text classification, as well as multi-task learning with transformers like BERT. It boasts a robust training system, easy model packaging, deployment, and workflow management, making it suitable for industrial-strength applications. spaCy is released under the MIT license, offering a comprehensive solution for developers and researchers working with NLP.

info8010-deep-learning is a GitHub repository offering a comprehensive set of lecture materials for the INFO8010 Deep Learning course at ULiège. This resource includes lecture PDFs, associated code examples (e.g., for polynomial regression, multi-layer perceptrons, automatic differentiation in PyTorch, convolutional neural networks, attention, transformers, GPT, graph neural networks, uncertainty, auto-encoders, and diffusion models), and homework assignments designed to familiarize users with the PyTorch library. It also provides a course syllabus, project guidelines, and archived lectures from previous editions, making it an invaluable educational tool for anyone looking to learn or teach deep learning concepts.

Stanford CME 295 Transformers & Large Language Models offers a comprehensive VIP cheatsheet for the Stanford CME 295 course. This resource condenses key concepts related to Transformers and Large Language Models, including self-attention mechanisms, architectural variants, and optimization techniques like sparse attention and flash attention. It also covers LLM-specific topics such as prompting, fine-tuning (SFT, LoRA), preference tuning, and optimization methods like mixture of experts, distillation, and quantization. The cheatsheet extends to practical applications like LLM-as-a-judge, RAG, agents, and reasoning models, making it an invaluable study aid for students and researchers.

Professor Chucky is an advanced real-time AI teaching platform designed to provide personalized learning experiences for students of all levels. It offers AI-powered tutoring in any subject and language, allowing users to learn at their own pace. Key features include transforming notes into interactive quizzes and AI-powered summaries, with options for custom quiz generation and ELI10 explanations. The platform also facilitates study groups for peer-to-peer learning and offers a magical storytelling adventure for kids to boost reading skills and creativity. Additionally, it includes an educational game called Pots & Pans to sharpen deduction skills, making learning interactive and fun.

aivancity School of AI & Data for Business & Society is a unique educational institution dedicated to Artificial Intelligence and Data Science, with campuses in Paris-Villejuif and Nice. It offers a range of programs from Bachelor's to Master's degrees, including a Grande École program, MSc in Data Engineering, Cloud Computing, Data Management, and AI for Business, as well as specialized MSc in Generative AI. The school emphasizes a hybrid approach combining AI, business management, and ethics, preparing students to become "IAgénieurs®" capable of addressing complex challenges in the economy and society. aivancity is recognized by the French state and aims to ensure long-term employability through its innovative "diploma update guarantee."

LearnEase is an AI-powered learning platform specifically designed to help students master the CBSE curriculum for grades 8-12. It provides comprehensive study materials, interactive quizzes, and personalized learning assistance to make complex topics more accessible. The platform leverages artificial intelligence to offer tailored explanations and practical examples, enhancing understanding and retention. LearnEase aims to simplify the learning process, allowing students to effectively prepare for their exams and improve their academic performance through engaging and adaptive content.

Blubridge operates as an independent AI research laboratory, dedicated to engineering advanced deep learning systems from foundational principles. The company specializes in providing enterprise-grade AI models, robust infrastructure, and comprehensive deployment solutions tailored for various business needs. Their expertise spans optimizing tokenization methods and refining AI systems to suit diverse applications, ensuring high performance and scalability. Blubridge focuses on modeling Artificial Intelligence Systems and their practical applications, offering cutting-edge solutions for organizations looking to leverage frontier AI research.

LLMRiddles is an open-source project that provides a game-like environment for users to explore and understand prompt engineering. Players are challenged to craft questions that interact with various language models (like ChatGPT, ChatGLM, DeepSeek, and Mistral-7B) to achieve specific, required outputs. This platform aims to deepen participants' understanding of how to cleverly construct prompts and trigger surprising responses from AI systems, highlighting the power of deep learning and natural language processing. It offers online versions for direct access and local deployment options, supporting multiple LLMs and languages. Users can also contribute custom levels to the game.

LLMsPracticalGuide is a comprehensive, actively updated resource offering a curated list of practical guides for Large Language Models (LLMs). It is based on a survey paper titled "Harnessing the Power of LLMs in Practice: A Survey on ChatGPT and Beyond" and includes an evolutionary tree of modern LLMs. The guide aims to assist practitioners in understanding and applying LLMs in natural language processing (NLP) applications, covering various aspects such as model types (BERT-style, GPT-style), data considerations (pretraining, finetuning, test data), NLP tasks (NLU, generation, knowledge-intensive), and practical concerns like efficiency, trustworthiness, and alignment. It also details usage restrictions based on model and data licensing information, making it a valuable resource for both research and practical implementation.

BetterBrain specializes in providing mid-market companies with production-ready AI solutions quickly, leveraging proprietary accelerators and a stack-agnostic approach. The platform offers full-stack delivery, covering everything from initial discovery and strategy mapping to development, deployment, adoption, and continuous optimization. Key offerings include BetterSearch for enterprise knowledge retrieval, BetterDocs for document intelligence, BetterAgent for custom AI agents, BetterVoice for voice agent automation, BetterChat for conversational AI, and BetterInsight for predictive analytics. BetterBrain aims to help companies transition from being AI-ready to AI-first, addressing common challenges like slow implementation times and pilot purgatory.

Autogen Tutorials is a Hugging Face Space dedicated to providing educational resources for AI automation using AutoGen. This platform helps users understand how to leverage multiple conversing AI agents to tackle complex tasks. The tutorials guide users through generating solutions, which can involve integrating various tools and human input. It's designed for individuals interested in exploring and implementing advanced AI automation techniques, offering practical insights into how AutoGen facilitates collaborative AI agent workflows for diverse applications.

The UKRI AI Centre for Doctoral Training (CDT) in Designing Responsible Natural Language Processing, based at the University of Edinburgh, focuses on training future researchers and innovators. The program emphasizes the development of responsible and trustworthy natural language processing systems. Studentships are available for PhD candidates, with applications opening for September 2026. The CDT covers five core skills domains, including responsible NLP data and models, explainable NLP, human-NLP partnership design, NLP governance and accountability, and co-creation for NLP futures. The center collaborates with a diverse community of partners from industry, advocacy groups, and policy-making organizations, offering students project challenges, placements, and internships.

Gooru Learning offers MyGooru AI for Personalized Pathways (MAP), an AI-driven personalization infrastructure designed to deliver assured outcomes across various industries including learning, finance, health, and enterprise. MAP goes beyond generative AI by using formal reasoning to build beliefs about each user and generate mathematically certain pathways, ensuring engagement and completion. It actively senses user mindsets, motivation, confidence, and intent, continuously updating probabilistic beliefs across knowledge, mindsets, interests, abilities, and community. This approach helps lower customer acquisition costs through personalized discovery and smarter conversion funnels, while increasing lifetime value via adaptive engagement and outcome completion. Gooru also provides tools for instructors, institution leaders, and curriculum developers.

Machine-Learning-Tutorials is an open-source GitHub repository offering a comprehensive, topic-wise curated list of machine learning and deep learning tutorials, articles, and other educational resources. It covers a wide range of subjects from foundational concepts like linear and logistic regression, model validation, and statistics, to advanced topics such as deep learning frameworks, natural language processing, computer vision, and reinforcement learning. The repository includes links to university courses, useful blogs, interview resources, and cheat sheets, making it a valuable hub for anyone looking to learn or deepen their understanding of AI and ML. It also features curated lists of R and Python tutorials specifically for data science, NLP, and machine learning.

Machine-Learning-Guide is an extensive open-source resource designed to help individuals learn about and improve their efficiency in machine learning development. It offers a wealth of information on various machine learning tools, libraries, frameworks, and large language models (LLMs). The guide includes learning resources, developer resources, courses, certifications, books, and YouTube tutorials. It also details specific ML/Deep Learning frameworks like TensorFlow, PyTorch, and Keras, along with tools for deploying and running LLMs. This resource is ideal for anyone looking to deepen their understanding of machine learning applications and development practices.