Research & Education

AIRS, or Artificial Intelligence Research for Science, is an open-source initiative offering a comprehensive collection of software tools, datasets, and benchmarks. It is specifically designed to support research in AI for quantum mechanics, density functional theory, small molecules, protein science, materials science, molecular interactions, biological science, partial differential equations, and ordinary differential equations. The project's goal is to foster an integrated, open, reproducible, and sustainable set of resources to advance the emerging field of AI for Science. It includes various methods and resources, with the list continuously expanding as research progresses, making it a valuable resource for academic and scientific communities.

async-rl offers a practical implementation of asynchronous 1-step Q learning, as detailed in the paper "Asynchronous Methods for Deep Reinforcement Learning." This open-source project leverages Tensorflow and Keras for deep Q network definition and optimization, while integrating with OpenAI's Gym library for interaction with the Atari Learning Environment. A key feature is its use of multiple actor-learner threads, which helps stabilize the learning process without relying on memory-intensive experience replay, making it efficient for machines with less RAM. The repository also includes a work-in-progress asynchronous advantage actor-critic implementation and provides instructions for training, visualizing with TensorBoard, and evaluating models.

Awesome-3D-Object-Detection is a comprehensive curated list of resources dedicated to deep learning for 3D Object Detection, with a strong emphasis on lidar-based methodologies. This GitHub repository serves as an invaluable hub for researchers and engineers, providing direct links to relevant academic papers, associated code implementations, and essential datasets like KITTI, nuScenes, Lyft, and Waymo Open Dataset. It also highlights top conferences and workshops in the field, offering a structured overview of the latest developments and trends. The resource includes surveys, books, videos, and course materials, making it a one-stop reference for anyone looking to delve into or stay current with 3D object detection.

awesome-deep-learning is an extensive, curated list designed for anyone interested in Deep Learning. It serves as a central hub for discovering high-quality educational materials and community resources. The list encompasses a wide array of content, from foundational books and university courses to insightful videos, academic papers, and practical tutorials. It also highlights researchers, relevant websites, datasets, conferences, and various frameworks and tools, making it an invaluable resource for both beginners and experienced practitioners looking to deepen their understanding or find specific information within the Deep Learning domain.

10x Science provides AI-native software specifically designed for scientists working with protein therapeutics. The platform significantly upgrades peptide mapping, offering processing speeds that are 10 times faster than traditional methods. It is engineered to uncover insights that existing tools might overlook, enhancing the depth and accuracy of protein characterization. A key differentiator is its ability to eliminate the need for file conversions, streamlining the workflow and reducing friction for researchers. This focus on speed, enhanced discovery, and user-friendly integration makes 10x Science a powerful tool for advanced scientific research.

3DAudio-Spectrum-Analyzer is an application designed for real-time visualization of audio spectra in a 3D environment. This tool allows users to generate binaural beats, offering a unique auditory experience. Key functionalities include the ability to start and stop audio analysis, calibrate devices for optimal performance, and precisely adjust the frequencies of the binaural beats. Hosted on Hugging Face Spaces, it provides an accessible platform for exploring audio dynamics and sound manipulation. The application is suitable for individuals interested in sound visualization and experimental audio generation.

Falcon-H1-Tiny offers a series of compact language models designed to push the boundaries of AI capabilities at a small scale. These models are available on Hugging Face Spaces and are ideal for research and experimentation. Users can input prompts and receive generated responses from these lightweight but capable AI models, making them suitable for various applications including research paper analysis, data visualization, and the development of small-scale AI applications. The focus on models with 100 million parameters or less makes them particularly efficient and accessible for developers and researchers working with limited resources.

SandboxAQ is a B2B company that utilizes the compound effects of AI and advanced computing to tackle significant societal challenges. Their Large Quantitative Models (LQMs) are applied across diverse fields, including AI drug discovery, new chemicals and materials innovation, cybersecurity, navigation, and medical diagnostics. The platform offers technologies such as AI simulation, cryptography management for enhanced cybersecurity, and AI sensing for global organizations. SandboxAQ's approach focuses on quantitative AI, grounded in physics and chemistry, to provide real-world solutions with clear outputs and reduced uncertainty, making it scalable and proven for high-stakes domains.

SpaceKnow Inc. leverages a cutting-edge platform to convert satellite data into actionable intelligence. Its core offering, SpaceKnow Guardian, integrates data from various satellite providers and employs proprietary AI algorithms to extract valuable information from complex raw data. This enables faster, data-driven decisions for users. The platform also features automated monitoring and an early warning system, providing timely alerts and continuous surveillance to detect changes and potential issues swiftly. SpaceKnow's solutions are tailored for sectors like Defense & Intelligence, enhancing situational awareness and strategic decision-making, and Construction Monitoring, offering precise tracking of progress and site activity.

WhiteLab Genomics is an AI-driven platform founded in 2019, backed by Y-Combinator, dedicated to accelerating the development of life-saving genomic medicines. The platform utilizes its proprietary AI-led framework, ALFRED (AI-Led Framework for Rational Exploration in Drug Design), which combines advanced algorithms, curated databases, and cutting-edge computational biology. This technology optimizes therapeutic candidates across diverse modalities such as AAV, lentivirus, and nanoparticles, focusing on safety and efficacy. WhiteLab Genomics supports various applications including gene and RNA-based therapies, target receptor identification, viral vector design, non-viral vector design, payload design, and bioproduction optimization. It aims to significantly accelerate and de-risk drug development processes.

Buluttan is an AI-based hyper-local weather intelligence platform designed to enhance operations, safeguard assets, and protect people with precision. The tool optimizes forecast algorithms and data flow to improve accuracy and precision, leveraging a Zoomcast AI Weather Model trained precisely to specific locations. It offers tailored insights for various sectors, including renewable energy, where it provides accurate power generation forecasts for wind farms and solar plants. For aviation, Buluttan delivers hyper-localized forecasts for airports, ensuring safer take-offs and landings. It also caters to the logistics, mobility, and port operations industries, providing advanced weather insights to navigate planning efficiently and manage tasks effectively.

Geoskop is an advanced climate intelligence platform designed to help renewable energy companies and other industries manage climate-related risks and optimize operations. It utilizes proprietary algorithms and AI to generate highly accurate, validated long-range climate predictions, enabling confident, climate-ready investment decisions. The platform assists in assessing long-term climate impacts on renewable assets, improving day-to-day performance through accurate seasonal forecasts, and anticipating extreme climate events. Geoskop also supports regulatory compliance with standards like the EU taxonomy and IFRS S2 through its Sustax tool, providing factual and fair-priced climate insights for reporting.

FOURIER-Robotics GR-2 is a cutting-edge humanoid robot designed to push the boundaries of agility, precision, and perception. Built upon customer feedback, GR-2 integrates advanced hardware, design, and software enhancements. Its next-level hardware design includes integrated cabling for power and communication, resulting in concealed wires and a more compact form factor. The improved joint configuration simplifies debugging, reduces manufacturing costs, and enhances the robot's ability to transition from AI simulation to real-world applications. GR-2 features 12-DoF dexterous hands, doubling the dexterity of previous models, and is equipped with six array-type tactile sensors for real-time force sensing and object manipulation. Powered by seven types of distinct FSA actuators, including FSA 2.0 with peak torques exceeding 380 N.m, GR-2 achieves dynamic mobility and precise movements. The Fourier Toolkit provides developers with an upgraded Software Development Kit, offering easy access to pre-optimized modules via intuitive APIs and supporting frameworks like NVIDIA Isaac Lab, ROS, and Mujoco.

Pluto Bio offers a collaborative multi-omics platform designed to accelerate research and drug discovery. It provides a unified workspace for preclinical and translational strategy, enabling multi-site, interdisciplinary collaboration in real-time. The platform centralizes data visualization with a no-code canvas, allowing users to explore data and test scientific hypotheses quickly while maintaining end-to-end traceability. Pluto Bio supports a wide range of biological assays, including scRNA-seq, RNA-Seq, ChIP-seq, ATAC-seq, and Spatial Transcriptomics, with pipelines for custom assays. It helps organize experiments, plots, data, and files in a secure cloud environment, facilitating target identification, biomarker discovery, and mechanism tracking.

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.

deep-learning-models is a GitHub repository offering Keras code and pre-trained weights for several widely used deep learning models. This resource includes implementations for VGG16, VGG19, ResNet50, Inception v3, and a CRNN for music tagging. The architectures are designed to be compatible with both TensorFlow and Theano backends, automatically adapting to the image dimension ordering specified in your Keras configuration. Users can easily load pre-trained weights, such as 'imagenet' for image models or 'msd' for the music tagging model, which are automatically downloaded and cached locally. While this repository is deprecated in favor of `keras.applications`, it remains a valuable reference for understanding and utilizing these foundational models.

Deep-Learning-TensorFlow is a GitHub repository offering a collection of pre-built Deep Learning algorithms implemented with the TensorFlow library. This package is designed as a command-line utility, enabling users to quickly train and evaluate popular Deep Learning models. It can also serve as a benchmark or baseline for comparing custom models and datasets. The repository includes implementations for Convolutional Networks, Restricted Boltzmann Machines, Deep Belief Networks, Deep Autoencoders, Denoising Autoencoders, Stacked Denoising Autoencoders, and MultiLayer Perceptrons. It also supports Logistic Regression. The package can be installed via pip as 'yadlt' or by cloning the GitHub repository, and it features a scikit-learn-like interface for ease of use.

UADAMAGE is an AI and GIS company specializing in geospatial analytics for automatic damage monitoring. The platform leverages satellite and drone imagery alongside advanced computer vision to assess damage following war or natural disasters. It transforms these diverse data inputs into actionable insights, aiding governments, organizations, and partners in making data-driven decisions. UADAMAGE's core focus areas include infrastructure recovery, demining efforts, and environmental monitoring, providing critical information for post-disaster assessment and planning.

ModelingToolkit.jl is a high-performance symbolic-numeric computation framework designed for scientific computing and scientific machine learning within the Julia ecosystem. It allows users to define models at a high level, enabling symbolic preprocessing for analysis and enhancement. The tool can automatically generate optimized functions for model components, such as Jacobians and Hessians, and automatically sparsify and parallelize computations. It also applies automatic transformations, like index reduction, to simplify models for numerical solvers. ModelingToolkit.jl supports composing multiple ODE subsystems and simulating complex Differential-Algebraic Equations (DAEs), making it a powerful tool for advanced scientific modeling and simulation.

efficient-gnns is a comprehensive repository offering code and resources for developing scalable and efficient Graph Neural Networks (GNNs). It specifically focuses on knowledge distillation techniques, including novel approaches like Graph Contrastive Representation Distillation, to create resource-efficient GNNs. The repository benchmarks various distillation methods, such as Local Structure Preserving loss and Global Structure Preserving loss, alongside baselines like Logit-based KD. It supports research on large-scale, real-world graph datasets for tasks like graph classification on MOLHIV and node classification on ARXIV and MAG, providing installation and usage instructions for researchers and developers in the field.

eo-learn is an open-source Python framework designed to streamline Earth observation processing and machine learning tasks. It provides a collection of Python packages that facilitate seamless access and automated processing of spatio-temporal image sequences from satellite fleets like Copernicus and Landsat. The framework is modular, allowing users to define sequences of operations for tasks such as cloud masking, image co-registration, feature extraction, and classification. It acts as a bridge between remote sensing and the Python data science ecosystem, making advanced tools accessible to non-experts while bringing state-of-the-art machine learning capabilities to remote sensing professionals. eo-learn uses NumPy arrays for data handling and supports various functionalities through modules like core, coregistration, features, geometry, io, mask, ml-tools, and visualization.

fire-detection-cnn is an open-source project offering real-time fire detection in video imagery through experimentally defined convolutional neural network (CNN) architectures. Based on research from ICIP 2018 and ICMLA 2019, it provides models like FireNet, InceptionV1-OnFire, InceptionV3-OnFire, and InceptionV4-OnFire for binary fire detection and superpixel-based localization. The tool emphasizes reduced complexity for high accuracy and computational performance, achieving up to 17 fps processing. It supports Python 3.7.x, TensorFlow 1.15, TFLearn 0.3.2, and OpenCV 3.x/4.x, and includes scripts for downloading pre-trained models and datasets. Users can convert models to protocol buffer (.pb) and tflite formats for integration with other frameworks like OpenCV DNN.

GazeML is a deep learning framework built on Tensorflow, designed for training high-performance gaze estimation models. It provides a robust platform for researchers and developers to implement and test various gaze estimation algorithms. The framework currently integrates re-implementations of published algorithms such as ELG (Eye region Landmarks based Gaze Estimation) and DPG (Deep Pictorial Gaze Estimation). While it may work on various platforms, it has been specifically tested on Ubuntu 16.04. Users can install dependencies, acquire pre-trained weights, and run webcam demos, making it a practical tool for advancing research in eye tracking and human-computer interaction.

REP, or Reproducible Experiment Platform, is an ipython-based environment designed for conducting data-driven research with an emphasis on consistency and reproducibility. It provides a unified Python wrapper for several machine learning libraries, including Sklearn, XGBoost, and Theanets, allowing users to work with a consistent interface. Key features include parallel training of classifiers on clusters, classification/regression reports with interactive plots, and smart grid-search algorithms with parallel execution. REP also supports research versioning using Git and offers pluggable quality metrics for classification. It aims to extend scikit-learn by providing a better user experience and tools for meta-algorithm design, making it a valuable resource for data scientists and researchers.