Research & Education

mdlm is an open-source masked discrete diffusion language model (MDLM) that features a novel substitution-based parameterization. This approach simplifies the absorbing state diffusion loss to a mixture of classical masked language modeling losses, leading to state-of-the-art perplexity numbers on LM1B and OpenWebText among diffusion models. It also achieves competitive zero-shot perplexity with state-of-the-art autoregressive models on various datasets. The repository provides the MDLM framework, simplified loss calculation, baseline implementations, and efficient samplers that make MDLM significantly faster than existing diffusion models, including semi-autoregressive generation capabilities.

MiniWorld is a minimalistic 3D interior environment simulator specifically designed for reinforcement learning and robotics research. It allows users to simulate environments featuring rooms, doors, hallways, and various objects, making it suitable for tasks like training AI agents in office, home, or maze-like settings. Written 100% in Python, MiniWorld is easily modifiable and extensible, offering features such as few dependencies, good performance, lightweight design, and support for domain randomization for sim-to-real transfer. It also provides fully observable top-down views, depth map production, and the ability to display alphanumeric strings on walls. This project has been deprecated as of August 11, 2025, and is no longer receiving updates or support.

Machine-Learning-Books-With-Python is an open-source GitHub repository designed to assist individuals in mastering machine learning concepts using Python. It offers comprehensive chapter-by-chapter notes, practical exercises, and corresponding code implementations for a variety of machine learning books. This resource is ideal for students and developers looking to deepen their understanding and practical skills in machine learning. The repository aims to provide a structured learning path, allowing users to follow along with popular textbooks and apply their knowledge directly through coding examples and solutions. It serves as a valuable companion for self-study and academic courses.

deep-learning-uncertainty is an open-source repository dedicated to predictive uncertainty estimation in deep learning models. It offers a comprehensive literature survey, detailed paper reviews, and experimental setups for various baseline methods. The repository also includes a collection of implementations, making it a valuable resource for researchers and engineers. This tool is designed to help users understand, quantify, and improve the reliability of predictions made by deep learning models, addressing critical aspects of model trustworthiness and robustness. It serves as a central hub for exploring established and emerging techniques in uncertainty quantification.

faceID_beta is an open-source project available on GitHub that provides an implementation of iPhone X's FaceID technology. It leverages face embeddings and siamese networks, processing RGBD images for facial recognition. The project is primarily presented as a Jupyter Notebook file, with an automatically generated Python file also available. This makes it particularly suitable for developers and researchers interested in understanding and experimenting with advanced facial recognition techniques. The repository includes details on the implementation and encourages users to explore the notebook version for a clearer understanding of the code's structure and functionality.

DriveLM is an open-source project focused on advancing autonomous driving research through Graph Visual Question Answering (GVQA). It provides comprehensive datasets, DriveLM-Data, built upon nuScenes and CARLA, specifically designed for driving with language. The project also offers DriveLM-Agent, a VLM-based baseline approach for jointly performing GVQA and end-to-end driving. DriveLM serves as a main track in the CVPR 2024 Autonomous Driving Challenge, offering a baseline, test data, submission format, and evaluation pipeline. It addresses the community's challenges by providing a benchmark for driving with language, exploring embodied applications of LLMs/VLMs, and investigating closed-loop planning with language.

DRL is an open-source collection of educational resources focused on Deep Reinforcement Learning. Hosted on GitHub, it offers a comprehensive set of materials including detailed slides, informative lecture notes, and explanatory videos, many of which are in Chinese. The repository covers fundamental and advanced topics such as Value-Based Learning (Q-learning, Sarsa, Experience Replay), Policy-Based Learning (REINFORCE, A2C, TRPO), Actor-Critic Methods, and specialized areas like Multi-Agent Reinforcement Learning and Imitation Learning. It's an excellent resource for students and researchers looking to deepen their understanding of DRL concepts and algorithms.

Musicgen Negative Prompting is an AI tool hosted on Hugging Face Spaces, designed to enhance music generation through the use of negative prompts. This functionality allows users to define elements or characteristics they wish to exclude from the generated music, offering a refined level of control over the creative process. By specifying what the music should *not* sound like, users can more effectively steer the AI towards desired outcomes, making it a valuable resource for refining musical ideas and exploring new creative boundaries. The tool is currently experiencing a runtime error, preventing its full functionality.

mlcourse is a comprehensive collection of machine learning course materials hosted on GitHub, offering a wide range of resources for learning fundamental machine learning concepts and techniques. The repository includes detailed lectures, homework assignments with programming problems, and conceptual checks. Topics covered span supervised and unsupervised learning, model evaluation, regularization techniques like Lasso and Elastic Net, kernel methods, and Bayesian statistics. It also features discussions on advanced topics such as gradient boosting, backpropagation, and various optimization methods. The materials are suitable for students and aspiring data scientists looking to deepen their understanding of machine learning principles through practical exercises and theoretical explanations.

Pangea is a fully open multilingual multimodal LLM developed by NeuLab at LTI/CMU, supporting 39 languages. It is designed for research and development in multilingual AI, offering a simple interface for text translation. Users can input text, select source and target languages, and receive a translated version. The tool is available as a Hugging Face Space, making it accessible for experimentation and integration into various projects. Its open-source nature under the Apache 2.0 license encourages diverse language applications and collaborative development within the AI community.

Paper-List is an open-source GitHub repository curated by Yanjie Ze, offering a comprehensive collection of research papers across the domains of robotics, learning, and computer vision. The list is meticulously organized by publication year and conference, including prominent venues like RSS, CVPR, ICLR, NeurIPS, CoRL, ICCV, ICML, and SIGGRAPH. It features papers on topics such as humanoid robots, dexterous manipulation, 3D robot learning, and robot foundation models. The repository also highlights 'Best Papers' and 'Recent Random Papers,' providing direct links to arXiv preprints, official websites, and other resources, making it an invaluable resource for researchers and academics to track cutting-edge advancements in these fields.

pytorch_diffusion offers a PyTorch reimplementation of Denoising Diffusion Probabilistic Models, complete with checkpoints converted from the original TensorFlow implementation. This tool allows users to load diffusion models with pretrained weights for various datasets like CIFAR-10, LSUN-bedroom, LSUN-cat, and LSUN-church. It provides a quickstart guide for running a Streamlit demo, making it accessible for immediate use. Users can also instantiate and configure the U-Net model for denoising independently. The repository includes instructions for producing samples, evaluating results against TensorFlow models, and converting TensorFlow checkpoints to PyTorch, making it a comprehensive resource for researchers and developers working with diffusion models.

PINNpapers is a comprehensive, open-source repository maintained by the IDRL lab, dedicated to curating essential research papers on Physics-Informed Neural Networks (PINNs). Since PINNs have gained significant traction in scientific computing, this resource serves as a valuable collection of representative works in the field. The repository categorizes papers across various aspects of PINNs, including foundational models, parallel computing approaches, acceleration techniques, model transfer and meta-learning, probabilistic PINNs, uncertainty quantification, and diverse applications. It also lists relevant software libraries like DeepXDE and SciANN, providing links to papers and code where available. Researchers and practitioners can use this resource to stay updated on the latest advancements and foundational concepts in PINN research.

pysc2-examples offers a collection of Deep Reinforcement Learning examples specifically designed for StarCraft II. Built upon Deepmind's pysc2, OpenAI's baselines, and Blizzard's s2client-proto, it provides a robust framework for developers and researchers. The project leverages TensorFlow 1.3 and includes examples for tasks like 'CollectMineralShards' using Deep Q Networks and A2C algorithms. Users can quickly set up the environment, install necessary libraries like pysc2 and baselines, download StarCraft II maps, and then train and enjoy their AI agents. It supports various parameters for training, including algorithm choice (deepq, a2c), total timesteps, exploration fraction, and options for prioritized replay or dueling networks.

BioSketch is a free, user-friendly drag-and-drop tool specifically designed for scientists and researchers to create high-quality scientific illustrations. It simplifies the process of building publication-ready figures by providing an extensive library of biomedical icons. Users can easily assemble complex diagrams and visuals without needing advanced graphic design skills. The platform aims to empower the scientific community to produce stunning and accurate visual representations of their research, enhancing communication and impact in academic publications and presentations.

Sloth is an open-source tool specifically designed for labeling image and video data, primarily catering to the needs of computer vision research. It enables researchers and data scientists to efficiently annotate visual data, which is crucial for training machine learning models. The tool supports various annotation tasks, making it a versatile solution for creating high-quality labeled datasets. Its open-source nature means it can be freely used and adapted by the community, fostering collaboration and customization in computer vision projects. Sloth aims to simplify the often complex and time-consuming process of data annotation, facilitating the development of robust AI applications.

Setup-NVIDIA-GPU-for-Deep-Learning is a comprehensive, open-source guide designed to assist users in setting up their NVIDIA GPUs for deep learning tasks. It outlines a clear, step-by-step process, starting with the installation of the latest NVIDIA GPU drivers. The guide then proceeds to cover essential software components such as Visual Studio with C++ support, Anaconda/Miniconda for package management, the CUDA Toolkit, and cuDNN. Finally, it provides instructions for installing PyTorch and includes a script to test the GPU setup, ensuring all components are correctly configured for optimal deep learning performance. This resource is invaluable for deep learning practitioners and AI researchers looking to streamline their development environment setup.

Stanford-CS-221-Artificial-Intelligence is a comprehensive resource offering VIP cheatsheets for Stanford's CS 221 Artificial Intelligence course. This repository aims to consolidate all crucial notions covered in the course, including cheatsheets for each artificial intelligence field and an ultimate compilation of concepts. The material is accessible on a dedicated website, ensuring readability across various devices. Authored by Afshine Amidi and Shervine Amidi, it serves as an invaluable study aid for students and anyone interested in understanding core AI principles. The cheatsheets are available in English, French, and Turkish, making it accessible to a broader audience.

tf-gnn-samples is a GitHub repository offering TensorFlow implementations of various Graph Neural Network (GNN) architectures. It serves as the code release for an article introducing GNNs with feature-wise linear modulation (GNN-FiLM). The repository includes implementations for Gated Graph Neural Networks (GGNN), Relational Graph Convolutional Networks (RGCN), Relational Graph Attention Networks (RGAT), Relational Graph Isomorphism Networks (RGIN), GNN-Edge-MLP, and Relational Graph Dynamic Convolution Networks (RGDCN). It provides scripts for training and evaluating models on tasks such as citation networks (Cora, Pubmed, Citeseer), protein-protein interaction (PPI), quantum chemistry prediction (QM9), and variable misuse detection (VarMisuse). The code allows users to reproduce experimental results presented in the accompanying research paper, making it a valuable resource for researchers and developers working with GNNs.

Abzu is a biotechnology company leveraging explainable AI to innovate in the field of RNA therapeutics. The company specializes in developing best-in-class RNA drugs, including siRNAs, ASOs, and anti-miRs, for significant medical needs. Their AI-guided design platform, powered by the QLattice®, allows for the in silico exploration and prioritization of vast sequence spaces, evaluating over 100,000 design variants to predict efficacy and developability properties. This approach significantly reduces experimental cycles, lowers costs, and shortens the time to candidate selection. Abzu also focuses on RNA-based delivery systems, developing targeted aptamers for cell-specific uptake of therapeutic RNA, offering a modular platform for precision delivery beyond the liver. The team combines deep RNA biology, AI-driven design, and drug development experience to create a closed learning loop where data refines models and models improve molecules.

Tensorflow-triplet-loss offers a robust implementation of triplet loss within the TensorFlow framework, specifically designed for metric learning tasks. It includes online triplet mining capabilities, which are crucial for training models that learn meaningful embeddings. The repository provides two main versions: "batch all" and "batch hard" triplet loss, allowing flexibility in how triplets are selected and processed. The code structure is adapted from CS230 assignments and is accompanied by tutorials, making it accessible for developers and researchers. It supports both CPU and GPU installations and includes scripts for training on datasets like MNIST, visualizing embeddings, and hyperparameter searching. This tool is ideal for those looking to implement or experiment with triplet loss for tasks such as face recognition or person re-identification.

Text2Human is an official PyTorch implementation for text-driven controllable human image generation, as presented in the SIGGRAPH 2022 paper. This open-source tool enables users to create human images by providing text descriptions that specify clothing shapes and textures. It includes a comprehensive framework for training and sampling, utilizing a large-scale, high-quality DeepFashion-MultiModal Dataset with rich multi-modal annotations. Researchers and developers can leverage its capabilities for tasks like generating images from parsing maps or human poses, and it offers a user interface for interactive text-to-human image generation. The project also provides pretrained models and detailed installation instructions, making it a valuable resource for AI research in computer graphics.

StudyX is an all-in-one AI study platform designed to enhance learning for students, educators, and professionals. It offers a comprehensive suite of AI-powered tools including homework help with step-by-step solutions, AI note-taking from various materials like PDFs and videos, and AI flashcard and quiz generators for effective exam preparation. Additionally, StudyX provides AI writing tools such as an AI detector, humanizer, plagiarism checker, and paraphraser to ensure originality and improve writing quality. The platform supports over 50 subjects and caters to different learning stages, from middle school to professional certification prep, making it a versatile resource for academic and professional development.

Score Jacobian Chaining is a technique designed for analyzing the sensitivity of machine learning models. This tool is invaluable for AI researchers and machine learning engineers seeking to understand the intricate relationship between model inputs and outputs. By providing insights into how changes in input data propagate through a model, it facilitates effective debugging and optimization. This understanding is crucial for improving model performance, ensuring robustness, and gaining deeper insights into model behavior. While the current live website indicates a runtime error, the underlying concept is highly relevant for academic research and practical application in machine learning development.