Research & Education

The Artificial Intelligence International Institute (AIII) is a Singapore-based AI think tank dedicated to promoting sustainable artificial intelligence for humanity. Established in 2017, AIII focuses on three core pillars: technology, commercialization, and governance, aiming to balance economic development, autonomy, governance, and ethics in AI evolution. The institute conducts research on autonomous enterprise transformation, fusionovation for venture creation, AI governance and risk management, and next-generation AI. AIII actively collaborates with various organizations and hosts events to foster interdisciplinary collaboration and responsible AI development.

Whattocode is an AI-powered platform specifically designed to generate frontend coding challenges. This tool aims to provide developers and coding students with a consistent and effective way to practice and improve their frontend development skills. By offering tailored exercises, Whattocode helps users enhance their abilities in various frontend technologies and concepts. The platform focuses on practical application, allowing users to engage with real-world coding scenarios to solidify their understanding and proficiency. It serves as a valuable resource for anyone looking to maintain or advance their frontend coding expertise through regular, targeted practice.

There.do is an AI tool designed for architects, engineers, and consultants to streamline the creation of site reports and meeting minutes. It allows users to capture notes and photos on-site, which are then automatically organized and integrated into documents. The AI assists with summarizing voice notes, rephrasing text for clarity, and proofreading. It also generates email drafts for sending reports and tracks who has read the documents, with options for reminders. This platform aims to replace traditional word processors by offering a continuous workflow from field notes to a sent, tracked document, significantly reducing the time spent on report generation.

RAGHub serves as a comprehensive, community-driven directory for the rapidly expanding field of Retrieval-Augmented Generation (RAG). It curates a living collection of new and emerging RAG frameworks, projects, and resources, addressing the challenge of keeping up with the constant influx of new tools. The platform aims to help users navigate the RAG ecosystem, providing a centralized place to discover innovations and assess the relevance of various tools. RAGHub categorizes resources into RAG Frameworks, Evaluation and Optimization Frameworks, Engines, Data Preparation Frameworks, Projects, and general Resources. It encourages community contributions, allowing users to add new tools and insights, fostering a collaborative environment for RAG development.

history-llms is an information hub for a project focused on training the largest possible historical Large Language Models (LLMs). These models are designed to be fully time-locked, meaning they only access information up to a specific knowledge-cutoff date, such as 1913, 1929, or 1946. This approach allows researchers to explore historical discourse patterns without the hindsight contamination present in modern LLMs. The project aims to provide tools for exploring massive textual corpora and complements traditional archival research, serving as a window into past perspectives on various topics. The models are intended for scientific applications, enabling research in humanities, social sciences, and computer science, with a commitment to minimizing interference with the normative judgments acquired during pretraining.

R-KV is a novel method for redundancy-aware KV cache compression specifically designed for large language models (LLMs) that rely on chain-of-thought (CoT) or self-reflection for reasoning tasks. It addresses the issue of bloated key-value (KV) caches during inference by ranking tokens on-the-fly for both importance and non-redundancy, retaining only the most informative and diverse ones. This approach allows for significant memory savings, up to 90%, and improved throughput (up to 6.6x) during long CoT generation, often with zero or even negative accuracy loss. R-KV is a plug-and-play, training-free solution that acts as a lightweight wrapper for any autoregressive LLM, making it easy to integrate into existing inference pipelines or RL roll-outs.

have-fun-with-machine-learning is an absolute beginner's guide to Machine Learning and Image Classification with Neural Networks, designed for programmers with no prior AI background. This hands-on guide focuses on practical application rather than theoretical concepts, showing users how to train a neural network to classify images. It covers setting up and using open-source technologies like Caffe and DIGITS, creating image datasets, training neural networks from scratch, and fine-tuning existing networks like AlexNet and GoogLeNet for improved accuracy. The guide emphasizes using existing tools in interesting ways to solve problems, making complex AI accessible without requiring deep mathematical understanding or powerful hardware.

snake-ga is an AI agent designed to learn how to play the classic Snake game from scratch using Deep Reinforcement Learning. The project leverages Deep Q-Learning, where the system receives state parameters and rewards based on its actions, gradually developing a strategy to maximize its score without explicit game rules. This approach enables the AI to achieve scores up to 50 points with a solid strategy after only five minutes of training. The tool also supports Bayesian Optimization to fine-tune the parameters of the Deep neural network and other Deep RL aspects. Implemented in Pytorch, it offers a robust platform for experimenting with AI in game environments.

qxresearch-event-1 is a GitHub repository providing a hands-on tutorial with over 50 Python applications, each meticulously crafted to be under 10 lines of code. This resource spans a wide array of topics including Machine Learning, Deep Learning, GUI development, Computer Vision, and API creation. Designed for both beginners and experienced developers, the concise nature of each application facilitates easy understanding and modification, making it an ideal platform for learning and experimenting with Python. The repository also offers video explanations for each project on the @qxresearch YouTube channel, enhancing the learning experience and allowing users to quickly grasp and customize the code. It fosters a community for Python enthusiasts to connect and stay updated on new projects.

SDXS provides real-time one-step latent diffusion models with image conditions, enabling rapid image generation. It boasts impressive inference speeds, generating 512x512 images at 100 FPS and 1024x1024 images at 30 FPS on a single GPU, making it 30x faster than SD v1.5 and 60x faster than SDXL for comparable image quality within a one-second generation limit. The tool also supports training ControlNet, expanding its applications to image-conditioned control and efficient image-to-image translation. SDXS utilizes a lightweight image decoder and a block removal distillation strategy for model acceleration, alongside a feature matching loss for efficient one-step model finetuning.

Secant is an AI-powered educational tool designed to enhance learning efficiency for both students and educators. It offers 24/7 AI support, enabling students to get assistance whenever needed, and provides integrated AI-powered learning tools to study smarter. For teachers, Secant automates various tasks, streamlining their workflow and allowing them to focus more on teaching. The platform aims to boost grades for students and is trusted by institutions. Secant has recently joined Teachshare to further power educational tools globally, indicating a commitment to providing robust solutions for the academic community.

BUZZcenter – AI dla firm specializes in practical AI training, consulting, and implementation services tailored for businesses. They offer a range of programs, including AI workshops, open training sessions, and an online learning platform called BUZZlearning, covering topics like AI in marketing, sales, HR, and data analysis. The company also provides individual mentoring, advisory services, and assistance with implementing AI tools like ChatGPT. Their approach focuses on delivering tangible results, helping companies increase efficiency, reduce costs, and accelerate processes by integrating AI into their operations. BUZZcenter emphasizes customized solutions, adapting programs to specific company needs and skill levels, and has been operating since 2018.

Self-Refine is an innovative AI research tool designed to empower Large Language Models (LLMs) with the ability to self-correct and enhance their output. The core mechanism involves LLMs generating feedback on their initial work, using this feedback to refine the output, and repeating this process iteratively. This iterative refinement process leads to improved quality and accuracy across various tasks. The tool provides examples and setups for diverse applications, including acronym generation, dialogue response generation, code readability improvement, and tasks like Commongen, GSM-8k, and Yelp. It utilizes 'prompt-lib' for querying LLMs and offers distinct prompt types for initialization, feedback generation, and iteration, making it a versatile platform for exploring self-improving AI systems.

java-go-python is an open-source GitHub repository offering an extensive collection of IT learning video tutorials. The repository is continuously updated and includes courses on popular programming languages like Java, Python, Go, C, C++, and C#. Additionally, it features video tutorials on front-end development, databases, big data, artificial intelligence, AIGC, ChatGPT, software testing, network security, reverse engineering, HarmonyOS application development, and Android. It serves as a valuable resource for developers, students, and IT professionals seeking to enhance their skills and stay current with the latest technologies.

self-critical.pytorch provides a comprehensive codebase for image captioning research, offering an unofficial PyTorch implementation for Self-critical Sequence Training. Key features include support for bottom-up features, test-time ensemble, and multi-GPU training, with DistributedDataParallel now supported via pytorch-lightning. The codebase also integrates Transformer captioning models and offers a simple demo via a Colab notebook. Researchers can train networks on datasets like COCO and Flickr30k, with options for scheduled sampling and evaluation using metrics like BLEU, METEOR, and CIDEr. Pretrained models are available, and the tool facilitates generating image captions and evaluating them on various splits.

ILearnDeepLearning.py is an open-source repository offering a collection of small projects focused on neural networks and deep learning. It serves as a practical companion to articles published on Medium, allowing users to explore the underlying mathematics and implementations of deep learning concepts. The repository includes projects on topics such as visualizing gradient descent, implementing neural networks with NumPy, preventing overfitting, and optimizing training processes. It also features examples of creating animated graphs and understanding convolutional neural networks. This resource is designed to help both beginners and those with some experience to deepen their understanding of deep learning through hands-on coding and visual explanations.

Baasin provides comprehensive AI education through its 'Baas in AI Academy,' offering online video courses like the ChatGPT Video Cursus. The platform also features an AI Shop for various AI-related products and a community, 'De AI Vereniging,' for networking and shared learning. Baasin caters to both individuals and businesses, with specific enterprise contracts available for teams, including options for email domain-based access and SSO. Beyond digital products, Baasin offers onsite AI workshops and a whitelabel workshop option, allowing organizations to deliver their own AI training using Baasin's materials. The platform emphasizes practical, immediately applicable knowledge to help users leverage AI for a competitive edge.

Hellbender Inc. specializes in crafting cutting-edge Computer Vision solutions, offering advanced AI vision systems and industrial AI cameras. They provide mission-critical hardware and software infrastructure for AI-driven perception systems, engineered for the edge in autonomy, robotics, and industrial applications. Their services include design, development, and turn-key manufacturing, with a focus on producing high-quality hardware in America. Hellbender also offers Computer Vision as a Service (CVaaS) for bespoke systems, addressing complex problems. They are a Raspberry Pi Design Partner and emphasize their commitment to employees, community, and the environment.

learning-papers is a curated collection of landmark papers in machine learning, designed to highlight important techniques and foundational research. The repository categorizes papers by topic, such as Deep Learning, Ensemble Methods, Optimization, and Natural Language Processing, making it easier to navigate significant contributions. Each entry often includes the paper's title, authors, publication year, and links to the paper itself, sometimes with alternative free versions or associated code. It also provides icons to indicate paywalled papers, freely available versions, associated code, precursor papers, iterations, blog posts, websites, videos, or slides, offering a comprehensive resource for understanding the evolution of machine learning concepts.

TransformerLens is an open-source Python library designed for the mechanistic interpretability of GPT-2 style language models. Maintained by Bryce Meyer and created by Neel Nanda, this tool enables users to load over 50 different open-source language models and expose their internal activations. Researchers can cache any internal activation and add functions to edit, remove, or replace these activations during model execution. The library supports in-depth analysis to reverse engineer the algorithms models learn from their weights, making it a crucial resource for understanding how large language models function internally. It also includes experimental support for Mamba / SSM architectures, providing bridge adapters for Mamba-1 and Mamba-2.

AntsBees is an AI training and solutions integrator specializing in driving talent in AI, Big Data, and Software Innovation. They provide comprehensive courses to help individuals and organizations stay ahead in the rapidly evolving digital landscape, covering topics like AI, cybersecurity, and emerging technologies. Beyond training, AntsBees delivers tailored software integration solutions designed to unite tools, automate workflows, and create smarter, more connected businesses. With over 13 years of industry experience, certified trainers, and a track record of serving leading organizations, AntsBees aims to bridge the gap between AI capabilities and the benefits organizations derive from them, facilitating digital transformation.

Tldr AI Summarizer is an intelligent reading companion designed to instantly summarize any article found on the web. This tool helps users save valuable time by providing concise summaries, allowing them to stay informed without sifting through lengthy content. It's particularly useful for cutting through clickbait and quickly grasping the main points of an article. Currently, Tldr AI is available as a web extension, with beta waitlists open for Chrome, Android, and macOS Safari versions, indicating future platform expansion.

Danti is an AI-powered search engine designed to help users make sense of the massive amounts of data collected globally. It allows users to simply ask questions, similar to a standard internet search, and then synthesizes data from various sources such as imagery, news, social media, and shipping information. This capability provides decision-making information quickly, regardless of the user's technical expertise. Danti offers multi-intelligence at your fingertips, enabling users to learn about any place on Earth through machine learning querying. It intelligently links related images, reports, and analysis, and is accessible via the web or deployable within organizational firewalls. Use cases include defense and intelligence, property and insurtech, and infrastructure, empowering users to access and understand complex data sets.

vecmap is an open-source framework designed to learn cross-lingual word embedding mappings. It enables users to build cross-lingual word embeddings from monolingual embeddings, with or without parallel data, using various methods including supervised, semi-supervised, identical, and fully unsupervised approaches. The framework also includes comprehensive evaluation tools for tasks such as word translation induction, word similarity/relatedness, and word analogy. It supports CUDA for faster processing on NVIDIA GPUs and is suitable for researchers and developers working on multilingual natural language processing tasks, particularly those focused on unsupervised machine translation.