Research & Education

LatentMAS is a multi-agent reasoning framework designed to enhance the efficiency and stability of multi-agent systems. Unlike traditional methods that rely on lengthy textual reasoning traces, LatentMAS facilitates agent collaboration by passing latent thoughts directly through their working memory within the model's latent space. This innovative approach significantly reduces token usage by 50-80% and achieves major wall-clock speedups of 3-7 times compared to standard Text-MAS or chain-of-thought baselines. The framework is compatible with any HuggingFace model and optionally supports vLLM backends for faster inference. It also features training-free latent-space alignment for stable generation, making it a general and powerful technique for developing advanced multi-agent AI applications.

2read is an AI-powered application designed to streamline the consumption of digital content by providing concise summaries and organizing information efficiently. It helps users quickly grasp key insights from articles, documents, and web pages, enhancing productivity and knowledge retention. The tool aims to simplify the process of understanding complex information by distilling it into easily digestible formats. This makes it ideal for anyone needing to process large volumes of text quickly, whether for academic research, professional development, or general knowledge acquisition. By focusing on efficient information organization, 2read supports better knowledge management and recall.

Stepwik offers a platform for discovering detailed codelabs and tutorials across a range of technologies. It aims to empower users on their coding journey by providing interactive learning experiences. The platform focuses on hands-on learning, allowing users to engage directly with the material rather than just reading static documentation. Stepwik is designed to support continuous learning and skill development in the technology sector, making complex topics more accessible through practical application.

LexLynk is a comprehensive Legal & Compliance tool designed to streamline legal research for professionals. It offers an interactive platform where users can navigate seamlessly through linked legal provisions, ensuring efficient research. Key features include a multi-column view for reading laws and texts side-by-side, direct linking of legal references, and an MS Office integration that allows users to access legal norms directly within Word and Outlook. The platform also provides a translation assistant for precise legal translations, collaborative functions for team annotations, and AI-supported analysis tools for comparing different versions of legal texts. Additionally, the LexTicker keeps users updated on current legal developments and changes.

SimWorx is an engineering, research, and development company founded in 2007 by postgraduate researchers from the State University of Campinas (UNICAMP). It specializes in innovative numerical modeling solutions, leveraging cutting-edge technology and AI to optimize oil, gas, and various engineering projects. SimWorx offers custom-built solutions tailored to specific industry needs, including computational vision, high-performance simulation, and AI-driven engineering. Key offerings include StimBR for acid stimulation analysis, WellWorx for production optimization, and Scope, an AI-based drilling NPT reduction tool. The company focuses on boosting performance and reducing costs through advanced simulation tools, enhancing engineering and decision-making processes for its clients.

FotonVR provides a comprehensive Virtual Reality in Education solution, specifically designed for K-12 schools. The platform offers a complete VR classroom setup, enabling immersive learning experiences across various subjects including Science, Mathematics, History, and STEM. With over 1200 curriculum-aligned VR modules, FotonVR aims to enhance student engagement and understanding. The content is suitable for grades up to 12, making it a versatile tool for different educational levels. FotonVR focuses on delivering an interactive and engaging learning environment through virtual reality technology, supporting educators in creating dynamic lessons.

Yippity is an AI-powered platform designed to streamline the creation of educational content. It allows users to convert any given text into various quiz formats, such as multiple-choice questions, true/false questions, and fill-in-the-blanks. Beyond quizzes, Yippity also generates study tools and flashcards automatically, aiming to help users learn faster and study smarter. The tool provides a straightforward way to transform notes and website content into interactive learning materials, making it suitable for educators, students, and anyone looking to create engaging study aids from existing text.

MathDash is an AI-powered math coaching platform specifically designed for students preparing for AMC 8, AMC 10, AMC 12, and MATHCOUNTS competitions. It offers a comprehensive training experience with over 10,000 practice problems, personalized training plans, and expert coaching to help users achieve success in math contests. The platform focuses on providing a structured approach to mastering complex mathematical concepts and improving problem-solving skills, making it an ideal resource for dedicated math competition participants.

AI_Challenger_2018 is an open-source platform designed to foster artificial intelligence talent globally by offering open datasets and programming competitions. It provides a structured environment for AI enthusiasts and professionals to test and develop their skills. The platform includes essential resources such as evaluation scripts for various competition tracks, enabling participants to accurately assess their model's performance. Additionally, baseline models are provided to help users kickstart their projects, offering a foundational understanding and a starting point for further improvements. While these baselines offer modest results, they are crucial for initial engagement and learning within the competition framework. The platform encourages continuous improvement and innovation in AI.

Yoodli AI Roleplays is an enterprise AI roleplay platform designed for communication coaching, sales enablement, partner training, and learning & development. It allows users to practice pitches, demos, and crucial conversations with AI coaches, receiving private, real-time, and judgment-free feedback. The platform offers unlimited reps with realistic AI simulations, breaking down performance with instant insights and allowing users to fine-tune specific aspects of their communication. Yoodli supports multi-persona roleplays for group presentations or interview panels and integrates with existing ecosystems for assigning roleplays, tracking progress, and syncing data. It is SOC 2 Type 2 certified and GDPR compliant, trusted by companies like Google and Snowflake.

Guide.AI is an innovative AI audio guide generator designed to help users create, publish, and monetize audio guides effortlessly. The platform leverages advanced AI text-to-speech technology, eliminating the need for manual audio recordings or professional translation services. This enables quick and easy creation of guides in multiple languages, boosting accessibility, inclusivity, and engagement for various audiences. Guide.AI aims to provide a free solution for content creators to generate passive income, making it an environmentally friendly and sustainable option for enhancing visitor experiences in sectors like travel, history, and culture.

Deep RL Course Certification is a specialized tool developed by Hugging Face to validate the completion of their Deep Reinforcement Learning Course. Users provide their Hugging Face username, first name, and last name to initiate the verification process. The application then assesses the user's model results to confirm successful completion of the course requirements. This certification serves as a valuable credential for individuals looking to demonstrate their proficiency in deep reinforcement learning, making it particularly useful for students and AI practitioners who have undertaken the Hugging Face course.

Listening is an AI text-to-speech tool designed specifically for academic papers and research. It converts PDFs, Word documents, MOBI & EPUB files, and even scanned physical pages into natural-sounding audio, allowing users to listen to complex material on the go. Key features include the ability to automatically skip citations, references, and footnotes, and to select specific sections of a paper to listen to. The tool also offers adjustable playback speeds and a one-click note-taking function that captures the last two sentences heard, timestamped and synced across devices. This helps students and researchers manage heavy reading loads, improve retention, and study more efficiently.

DeepLearningInMedicalImagingAndMedicalImageAnalysis is an open-source GitHub repository dedicated to curating resources for deep learning in medical imaging and medical image analysis. It serves as a valuable hub for researchers, developers, and practitioners interested in the intersection of AI and healthcare. The repository compiles reviews, surveys, and various materials that shed light on the current state and future potential of deep learning techniques in medical contexts. Users can explore a wide range of topics, from diagnostic support to image segmentation and reconstruction, gaining insights into how AI is transforming medical diagnostics and treatment planning. This resource is ideal for those looking to understand the theoretical foundations and practical applications of deep learning in this specialized field.

Easy Save AI is a comprehensive directory of digital marketing tools, including many powered by AI, designed to help users discover new solutions daily. Each tool listed is reviewed by Yusuf Muritala and his team, ensuring quality and relevance. The platform features a wide range of categories from content creation and marketing to analytics and SEO. Users can browse tools, filter by type, pricing model (free, paid, subscription, etc.), and features. Additionally, Easy Save AI offers a unique service to submit tools to over 100 directories, saving users significant time and boosting their online presence with high-DR backlinks.

deep-image-retrieval is an open-source project from Naver Labs Europe focused on advancing image retrieval through deep learning. It offers models and evaluation scripts implemented in Python3 and PyTorch 1.0+, enabling researchers and developers to learn deep visual representations for image retrieval tasks. The tool supports training image retrieval systems using various loss functions, including triplet loss and a novel Average Precision (AP) loss, which directly optimizes for retrieval performance. It includes pre-trained models based on Resnet architectures with different pooling mechanisms (MAC, GeM) and provides scripts for evaluating these models on standard benchmarks like Oxford5K and Paris6K, as well as for extracting features from custom image datasets.

iMemo AI Note Taking & Summary is an intelligent note-taking application designed to effortlessly capture and transcribe meetings, lectures, and podcasts. It records audio on the go and instantly transcribes it into text in over 100 languages. The tool leverages AI to summarize lengthy recordings, making information more digestible and easier to review. Users can organize notes into project folders and utilize an advanced search feature to quickly find specific information. iMemo also offers a dual-language feature for viewing content in two languages simultaneously and can translate notes into over 120 languages. It prioritizes user focus by not joining meetings with 'meeting bots', allowing users to stay present.

OWL Learning, powered by MPS, provides expert online learning solutions for all learners. With over 50 years of experience, OWL Learning partners with academic and industry partners to develop content and learning experiences for learners at any level and delivery in any modality on any platform. They offer a full suite of services including content development, learning design, creative media services, extended reality and gamification, translation, AI/ML services, and accessibility. OWL Learning delivers scaled solutions to meet development needs, whether for consultation to develop a vision or expertise to scale development.

Deep-Learning-Approach-for-Surface-Defect-Detection is an open-source project offering a Tensorflow implementation of a segmentation-based deep learning approach for surface defect detection. This tool is designed for automated visual inspection and quality control, particularly relevant in manufacturing processes. It allows users to train a deep learning model on datasets like KolektorSDD to identify and classify surface imperfections. The implementation supports independent training of segmentation and decision networks, providing flexibility for model optimization. It includes scripts for testing, training, and visualization of results, making it a practical resource for researchers and developers working on computer vision applications for industrial quality assurance.

LightNet is an open-source project offering a collection of light-weight neural networks specifically designed for semantic image segmentation. It focuses on achieving high segmentation accuracy while maintaining computational efficiency, making it suitable for embedded devices often found in autonomous driving systems. The repository includes implementations of several architectures such as MobileNetV2Plus, RF-MobileNetV2Plus, MobileNetV2Vortex, MobileNetV2Share, Mixed-scale DenseNet, SE-WResNetV2, and ShuffleNetPlus. These models incorporate techniques like Spatial-Channel Squeeze & Excitation (SCSE), Receptive Field Block (RFB), and Vortex Pooling. LightNet provides code in PyTorch and supports training and evaluation on Cityscapes and Mapillary Vistas Datasets, along with data augmentation using GANs.

Deep-Learning-Book-Chapter-Summaries is a GitHub repository dedicated to making the comprehensive Deep Learning book by Ian Goodfellow, Yoshua Bengio, and Aaron Courville more accessible. It provides detailed summaries for each chapter, breaking down complex topics into easier-to-understand explanations. The project also includes blog posts for particularly challenging chapters, offering further insights and elaborations. This resource is ideal for students and researchers looking to grasp the core concepts of deep learning without having to delve into every intricate detail of the original textbook, serving as a valuable study aid and reference.

Deep-Learning-for-Medical-Applications is a comprehensive, open-source repository on GitHub, providing a curated list of deep learning papers specifically focused on medical image analysis. This resource is designed to be a valuable starting point for researchers and practitioners in the field of medical AI. It meticulously classifies papers based on their deep learning techniques (e.g., CNN, RNN, GAN) and learning methodologies, along with metadata such as imaging modality, area of interest, and clinical database. The collection includes papers published since 2015 from peer-reviewed journals and high-reputed conferences, as well as recent arXiv preprints. The repository offers shortcuts to understand common deep learning techniques and imaging modalities, making it easier to navigate the extensive list of research.

Unlost provides a structured program designed to assist students in navigating their post-school journey. Through 6 weekly one-on-one sessions with a trained near-peer facilitator, participants develop a clear and actionable plan for their future. The program focuses on creating concrete post-school plans, identifying contingency pathways, and fostering the confidence needed to pursue these goals. While the website content is minimal, the core offering revolves around personalized guidance and support for students transitioning out of school, aiming to reduce uncertainty and empower them with a strategic outlook.

Touring Test is an innovative geography quiz game that leverages generative AI to create unique content for each round. Players are challenged to guess a city based on a variety of AI-generated clues, including poems, travel blogs, paintings, and descriptions of local food. The game offers flexible round times (15s, 30s, 90s) and varying numbers of rounds (5, 30) to suit different play styles. It supports both single-player and in-browser multiplayer modes, making it a fun and interactive experience for friends. The content is explicitly stated to be AI-generated, providing a novel way to engage with geography and AI creativity.