google-research/google-research

Features

• Quantum Computing - Develops superconducting and neutral atom qubits to achieve verifiable quantum advantages in scientific fields.
• Scientific Computing - Provides foundational computational frameworks and high-performance abstractions for complex scientific modeling and large-scale data analysis.
• Quantum Tomography Protocols - Shadow tomography estimates expectation values of quantum observables using sample-efficient and time-efficient measurement protocols that entangle constant numbers of state copies.
• Foundational Research Frameworks - Provides foundational algorithms, distributed training tools, and responsible AI practices for developing complex models.
• Experimental Research Platforms - Facilitates testing of experimental machine learning algorithms and computational graphs for advanced research.
• Responsible AI Frameworks - Implements frameworks for fairness, transparency, and robustness to ensure ethical standards in algorithmic development.
• Machine Learning for Science - Utilizes computational techniques to address large-scale challenges in fields such as health, sustainability, and crisis resilience.
• Computational Graphs - Executes machine learning models using computational graphs for automatic differentiation and gradient-based optimization.
• Quantum Research Toolkits - Provides a suite of simulation environments and algorithmic tools for exploring quantum physics principles.
• Quantum Simulators - Provides specialized quantum simulation environments for modeling algorithms beyond classical capabilities.
• Scientific Discovery Acceleration - Provides AI-powered tools to help scientists solve complex problems and accelerate the discovery process.
• Awesome List - A community-curated directory that catalogs and links out to other open-source projects, rather than a standalone tool you run yourself.
• Flood Risk - Processes environmental data through machine learning models to provide actionable flood prediction information.
• Public Health Trend - Generates high-accuracy forecasts for infectious disease hospitalizations by analyzing historical and real-time data.
• Weather Forecast Generation - Produces hyper-precise nowcasts and long-range climate predictions using neural network-based models.
• Frontier AI Models - Integrates frontier AI models to embed intelligent reasoning and generative capabilities into research workflows.
• Distributed Gradient Synchronization - Coordinates distributed parameter synchronization across compute nodes to enable large-scale parallel model training.
• Inference Acceleration - Reduces latency in autoregressive models by predicting and verifying multiple tokens in parallel.
• Machine Learning Training - Provides high-performance libraries and environments for training and rapid experimentation with complex models.
• Training Acceleration Engines - Accelerates large-scale model training using purpose-built hardware and cloud-based compute clusters.
• Hardware Acceleration - Optimizes high-performance machine learning workloads through hardware-accelerated tensor processing.
• Automatic Speech Recognition - Performs automatic speech recognition across hundreds of languages using large-scale, pre-trained models.
• Factuality Benchmarking Frameworks - Evaluates and improves the factual consistency of large language models through robust testing frameworks.
• Generative Document Ranking - Uses large language models to process queries and candidate documents to identify and order relevant information.
• Medical Information Provision - Delivers high-quality, evidence-based answers to medical inquiries using large language models fine-tuned for clinical domains.
• Genomic Sequence Interpreters - Applies deep learning to process and interpret genetic data for variant identification.
• Natural Disaster Prediction - Forecasts riverine and flash floods using hydrologic models and satellite-derived datasets to provide early warnings.
• Clinical AI Development Kits - Supplies pre-trained models and development tools to help engineers build specialized clinical applications.
• Oncology Imaging Analyzers - Analyzes medical imagery to identify early signs of malignancy, improving diagnostic accuracy.
• Quantum Error Decoding - Identifies likely error subsets in quantum circuits using heuristic-guided pathfinding algorithms.
• Systems Engineering - Provides foundational software and networking frameworks to support advanced research and quantum computing infrastructure.
• Natural Hazard Detection - Provides early warning systems for natural disasters like floods and wildfires using real-time environmental data.
• Foundational Algorithm Development - Advances core machine learning theory, information retrieval, and perception systems to solve complex computing challenges.
• River Flow Prediction - Processes geographical and meteorological data through neural networks to forecast river flow rates and flood risks.
• Privacy-Preserving Analytics - Extracts aggregate insights from sensitive datasets using cryptographic protocols and differential privacy.
• Privacy-Preserving Machine Learning - Provides libraries and tools to perform machine learning and data analysis while protecting individual privacy and sensitive information.
• Differential Privacy Aggregators - Extracts aggregate insights from sensitive datasets using differential privacy and synthetic data generation to maintain user trust.
• Private Data Processing Environments - Collects and combines sensitive information from multiple devices using secure, privacy-preserving cryptographic protocols.
• Algorithmic Fairness Auditing - Assesses machine learning algorithms for equitable outcomes and generalizability across diverse populations.
• System Robustness Analysis - Examines AI models for reliability and interpretability to ensure predictable performance in real-world applications.
• Forecasting - Generates high-confidence extreme weather forecasts using large-scale geospatial datasets and predictive modeling.
• Media Synthesis from Text - Generates high-fidelity images, videos, and audio from textual descriptions using large-scale generative models.
• Geospatial Health Analytics - Processes environmental and population data to predict disease outbreaks and identify local vulnerabilities.
• Planted - Applies quantum algorithms to solve noisy inference problems with quartic speedups.
• Perception Dataset Processors - Utilizes parallel computing clusters to train machine learning models on massive datasets for perception tasks.
• Edge AI Model Deployment - Optimizes and deploys machine learning models to run efficiently on local hardware and edge devices.
• Scientific Model Evaluators - Evaluates predictive models across diverse domains by comparing results against established datasets.
• ML Performance Profilers - Analyzes performance across CPUs and accelerators to provide actionable optimization suggestions for large-scale workloads.
• Multi-Agent Orchestration - Orchestrates multi-agent workflows to delegate complex search and analysis tasks across autonomous agents.
• Physics-Informed Architectures - Integrates physical laws into neural network architectures to improve simulation accuracy for complex systems.
• Visual Recognition Classifiers - Identifies and classifies objects within visual data using deep convolutional neural networks.
• Agentic Visual Reasoning - Composing zero-shot multimodal reasoning with language models.
• Automated Machine Learning - Automatically discover programs for ML tasks.
• Benchmarks and Datasets - Source for MBPP and other program synthesis evaluation datasets.
• Computer Vision - Collection of supervised contrastive learning and research implementations.
• Cross-Modal Models - Transformers for multimodal self-supervised learning from raw data.
• Density Functional Theory - Provides building blocks for differentiable DFT calculations.
• Discrete Diffusion Models - Implements structured denoising models in discrete state spaces.
• Domain Adaptation - Wasserstein distance-based batch effect correction.
• Embodied Perception - Collection of research datasets including visual perception tasks.
• Evaluation Metrics - Multi-scale image quality transformer implementation.
• General Purpose Models - Collection of research code including unified language learning model implementations.
• Machine Learning - Official repository for Google Research projects.
• Multimodal Alignment - Learning temporal cycle-consistency across multimodal sequences.
• Multimodal Transformers - Transformers for self-supervised learning from raw video, audio, and text.
• Natural Language Processing - Rethinking attention mechanisms with performer architectures.
• Neural Architecture Search - Regularized evolution for classifier architecture search.
• Neural Network Frameworks - Research code for gradient estimation and unrolled computation graphs.
• Neural Radiance Field Implementations - Official repository containing various research implementations and experiments.
• Neural Radiance Fields - Official research repository containing various JAX-based scene representation models.
• Open Source Models - Repository containing various research models including unified language paradigms.
• Reasoning and Perception - Language model programs for embodied control.
• Related Restoration Tasks - Synthesizes motion blur for training restoration models.
• Robust Learning Frameworks - Distills effective supervision from severe label noise.
• Sequence To Sequence Models - Repository containing instruction-tuned language model implementations.
• Data Processing and Analysis - Processes massive datasets across distributed systems to extract insights for scientific discovery.
• Geospatial Processing - Processes planetary imagery and satellite information using foundation models to generate actionable environmental insights.
• Similarity Search - Large-scale inference acceleration using anisotropic vector quantization.
• Body Models: - Listed in the “Body Models:” section of the Curated List Of Awesome 3D Morphable Model Software And Data awesome list.
• Automated Workflow Engines - Orchestrates autonomous multi-agent systems to perform complex coding and testing tasks in parallel.
• Computational Biology - Applies machine learning to decode genomic data and map neural connections for biological research.
• Content-Based Analysis Engines - Extracts meaningful information from images, audio, and video to enable content-based search and classification.
• Language and Robotics Integration - Combines natural language understanding with physical robot control to execute complex tasks in real-world environments.
• Environmental Modeling - Analyzes global geospatial data to track environmental changes and provide insights for ecological research.
• Disaster Response Coordination - Processes real-time data to assist in emergency management and rapid decision-making during natural disasters.
• Code Synthesis - Searches scientific literature to write and evaluate code solutions for complex computational problems.
• Mathematical Solution Derivation - Solves complex scientific equations by systematically exploring mathematical techniques and integrating symbolic reasoning.
• Neural Circuitry Modeling - Simulates and validates functional mechanisms in biological systems by testing circuit hypotheses against structural data.
• Research and Data Analysis Tools - Facilitates scientific data analysis through domain-specific models and pipelines for high-fidelity simulation.
• Traffic Flow Optimization - Analyzes vehicle movement patterns to provide intelligent signal timing recommendations that reduce congestion.
• Urban Infrastructure Mapping - Provides large-scale datasets of building footprints and temporal urban changes to support city planning.
• Cryptographic Execution Proofs - Provides hardware-based attestation to verify that data aggregation processes execute authorized code without tampering.
• Diagnostic Interview Simulators - Uses conversational diagnostic reasoning to assist clinicians in gathering patient history and performing assessments.
• Forecasting Benchmarks - Evaluates the performance of data-driven predictive models for weather, climate, and environmental events against standardized datasets.
• Hydrological Model Training - Utilizes historical river datasets and custom local data to train machine learning models for specific watershed environments.
• Predictive User Interfaces - Integrates machine learning to anticipate user intent and automate interface responses.
• Recommender Systems - Suggests relevant content and activities to users by analyzing engagement data within interactive environments.
• Video-based Pulse Estimators - Analyzes facial video clips using temporal shift neural networks to detect blood pulse fluctuations.
• Scientific Research Agents - Analyzes research data to propose new scientific insights and support experimental discovery.
• Retrieval Iteration Loops - Implements iterative retrieval loops to ensure sufficient context is gathered for accurate query responses.
• Engagement Analytics - Tracks and interprets interaction metrics to evaluate system performance and inform design improvements.
• Multi-Source Weather Data Integration - Combines diverse meteorological inputs from satellite and gauge-based sources into a unified framework to improve predictive accuracy.
• Research Datasets - Provides curated, benchmark-ready research datasets to support scientific validation and machine learning research.
• Research Dataset Documentation - Provides frameworks and templates to improve transparency and accountability in dataset creation.
• Data Extraction - Distills high-resolution atmospheric measurements from satellite imagery using physics-guided neural networks.
• Climate Impact Mitigation - Predicts the formation of heat-trapping condensation trails to enable real-time flight altitude adjustments.
• Geospatial Mapping - Provides comprehensive datasets of building locations to support urban planning and social good initiatives.
• Combinatorial Optimization Problems - Applies specialized algorithms to address complex combinatorial challenges in research and industrial applications.
• Iterative Feedback Loops - Refines programmatic solutions for scientific tasks by iteratively testing code through feedback loops to improve performance.
• Experimental Research Models - Provides access to experimental machine learning and artificial intelligence implementations for academic and technical research.
• Contextual Scope Enforcement - Facilitates privacy-first exchange of personal information between heterogeneous consumer devices using contextual scope enforcement.
• Modular Research Frameworks - Provides modular research abstractions to encapsulate algorithms and datasets into interchangeable components.
• Pipeline Performance Evaluators - Assesses the effectiveness of ranking models and AI assistants by analyzing potential biases and limitations in automated judgment processes.
• Network Control Validators - Detects and alerts operators to incorrect inputs within software-defined networking controllers.
• Predictive Text Inputs - Predicts and suggests text sequences to streamline user interaction and improve typing efficiency.
• Textual Entity Extractors - Identifies and categorizes people, organizations, and locations within unstructured text to support information extraction.
• Image Retrieval Systems - Optimizes image retrieval pipelines by combining global feature searching with local feature re-ranking for high-precision matching.
• Semantic Analysis Tools - Determines the intent and logical meaning of language data through classification and reasoning.
• Syntactic Parsers - Generates tree representations and dependency graphs to map the grammatical hierarchy of sentences.
• Cross-Source Querying - Provides cross-corpora query routing to identify and select relevant data sources for complex multi-step analysis.
• Physiological Signal Estimators - Aggregates periodic heart rate measurements using confidence scores and filtering to estimate resting heart rate.
• Compute Grants - Supplies researchers with free compute resources and infrastructure access to accelerate experimental research.
• Immersive and Interactive Systems - Constructs immersive user interfaces with dynamic simulations and interactive media to enhance engagement.
• Research and Analysis Tools - Maintains open source projects and programming interfaces to support machine learning and artificial intelligence research.
• Data Visualization - Transforms intricate datasets into interactive graphical representations to facilitate exploration and discovery.

Star history