EHR-Safe: Generating high-fidelity and privacy-preserving synthetic electronic health records
Differential privacy accounting by connecting the dots
Accelerating text generation with Confident Adaptive Language Modeling (CALM)
Who said what? Recorder's on-device solution for labeling speakers
RT-1: Robotics Transformer for real-world control at scale
Formation of robust bound states of interacting photons
Private Ads prediction with DP-SGD
Google at EMNLP 2022
Will You Find These Shortcuts?
Talking to Robots in Real Time
Making a Dual of a Traversable Wormhole with a Quantum Computer
Better Language Models Without Massive Compute
Google at NeurIPS 2022
Conversation Summaries in Google Chat
The Data Cards Playbook: A Toolkit for Transparency in Dataset Documentation
Mixture-of-Experts with Expert Choice Routing
Characterizing Emergent Phenomena in Large Language Models
Multi-layered Mapping of Brain Tissue via Segmentation Guided Contrastive Learning
ReAct: Synergizing Reasoning and Acting in Language Models
Infinite Nature: Generating 3D Flythroughs from Still Photos
Beyond Tabula Rasa: Reincarnating Reinforcement Learning
Robots That Write Their Own Code
Natural Language Assessment: A New Framework to Promote Education
Open Images V7 — Now Featuring Point Labels
Google at ECCV 2022
PI-ARS: Accelerating Evolution-Learned Visual-Locomotion with Predictive Information Representations
MUSIQ: Assessing Image Aesthetic and Technical Quality with Multi-scale Transformers
Do Modern ImageNet Classifiers Accurately Predict Perceptual Similarity?
Table Tennis: A Research Platform for Agile Robotics
UL2 20B: An Open Source Unified Language Learner
Crossmodal-3600 — Multilingual Reference Captions for Geographically Diverse Images
AudioLM: a Language Modeling Approach to Audio Generation
Large Motion Frame Interpolation
Quantization for Fast and Environmentally Sustainable Reinforcement Learning
TensorStore for High-Performance, Scalable Array Storage
View Synthesis with Transformers
FindIt: Generalized Object Localization with Natural Language Queries
Google at Interspeech 2022
Robust Online Allocation with Dual Mirror Descent
PaLI: Scaling Language-Image Learning in 100+ Languages
LOLNeRF: Learn from One Look
Learning to Walk in the Wild from Terrain Semantics
A Multi-Axis Approach for Vision Transformer and MLP Models
Digitizing Smell: Using Molecular Maps to Understand Odor
Announcing the Patent Phrase Similarity Dataset
High-Definition Segmentation in Google Meet
Using ML to Boost Engagement with a Maternal and Child Health Program in India
UVQ: Measuring YouTube's Perceptual Video Quality
OptFormer: Towards Universal Hyperparameter Optimization with Transformers
Towards Helpful Robots: Grounding Language in Robotic Affordances
Rax: Composable Learning-to-Rank Using JAX
Efficient Video-Text Learning with Iterative Co-tokenization
Introducing the Google Universal Image Embedding Challenge
Building Efficient Multiple Visual Domain Models with Multi-path Neural Architecture Search
Efficient Sequence Modeling for On-Device ML
Enhancing Backpropagation via Local Loss Optimization
Look and Talk: Natural Conversations with Google Assistant
ML-Enhanced Code Completion Improves Developer Productivity
Training Generalist Agents with Multi-Game Decision Transformers
Simplified Transfer Learning for Chest Radiography Model Development
Google at ICML 2022
Towards Reliability in Deep Learning Systems
Rewriting Image Captions for Visual Question Answering Data Creation
Revisiting Mask Transformer from a Clustering Perspective
Deep Hierarchical Planning from Pixels
Enabling Creative Expression with Concept Activation Vectors
MLGO: A Machine Learning Framework for Compiler Optimization
Identifying Disfluencies in Natural Speech
Minerva: Solving Quantitative Reasoning Problems with Language Models
Quantum Advantage in Learning from Experiments
Mapping Urban Trees Across North America with the Auto Arborist Dataset
Google at CVPR 2022
Scanned Objects by Google Research: A Dataset of 3D-Scanned Common Household Items
LIMoE: Learning Multiple Modalities with One Sparse Mixture-of-Experts Model
End-to-end Generative Pre-training for Multimodal Video Captioning
Deep Learning with Label Differential Privacy
Image-Text Pre-training with Contrastive Captioners
Vector-Quantized Image Modeling with Improved VQGAN
Contextual Rephrasing in Google Assistant
Challenges in Multi-objective Optimization for Automatic Wireless Network Planning
Language Models Perform Reasoning via Chain of Thought
Unlocking Zero-Resource Machine Translation to Support New Languages in Google Translate
Learning Locomotion Skills Safely in the Real World
GraphWorld: Advances in Graph Benchmarking
Alpa: Automated Model-Parallel Deep Learning
Extracting Skill-Centric State Abstractions from Value Functions
