Introducing our latest research on smartwatch-based walking metrics estimation. Historically, capturing precise gait metrics like double support time or swing time required specialized lab equipment. Our team has developed a multi-head temporal convolutional network that allows a single smartwatch to estimate these vital health biomarkers with high reliability. Through a study with 246 participants, we verified that smartwatch-based estimates are comparable to smartphone-based methods. This work is a significant step toward making longitudinal gait monitoring accessible for applications such as early disease detection, fall risk prevention, and personalized rehabilitation planning. Learn more: https://goo.gle/4pEQBfW
About us
From conducting fundamental research to influencing product development, our research teams have the opportunity to impact technology used by billions of people every day. We aspire to make discoveries that impact everyone, and sharing our research and tools to fuel progress in the field is fundamental to our approach.
- Website
-
https://research.google/
External link for Google Research
- Industry
- Technology, Information and Internet
- Company size
- 1,001-5,000 employees
Updates
-
Can chatbots serve doctors and patients? 🩺🤖 The role of chatbots in medicine is evolving fast. But how can they be most helpful? We curated a new notebook with Google for Health exploring the intersection of human-centered AI and patient journeys. It covers our research on helping physicians, patients, and developers. Check out the notebook here: https://lnkd.in/gS-7DBxt
-
-
Traditional traffic safety evaluation relies on historical crash data, which is often a "lagging" indicator and statistically sparse on many roads. Our Mobility AI team established a statistically significant correlation between hard-braking events (HBEs) and actual road segment crash rates. By analyzing anonymized data from Android Auto, we found that HBEs provide a signal 18 times more dense than traditional crash reports, allowing for the identification of high-risk locations — like the Highway 101 and 880 merge in California — without waiting for years of collision history. We are now working with Google Maps Platform to integrate these leading indicators into the Roads Management Insights offering to help transportation agencies improve infrastructure safety. Learn more: https://goo.gle/4bq92Sg
-
-
Today we announce the release of MedGemma 1.5, an update to our open medical AI models for developers. As healthcare industry AI adoption accelerates, we are expanding the Health AI Developer Foundations (HAI-DEF) collection to support more complex clinical workflows. MedGemma 1.5 4B improves baseline performance on high-dimensional 3D imaging — including CT and MRI — and substantially improves reasoning over medical records. Additionally, we recently introduced MedASR, our new automated speech recognition model fine-tuned for medical dictation that achieves up to 82% fewer errors on specialized medical benchmarks compared to generalist models. To encourage further innovation, we are also kicking off the MedGemma Impact Challenge on Kaggle with $100,000 in prizes. Learn more: https://goo.gle/4qVhQUR MedGemma Impact Challenge: https://lnkd.in/gYYnFFkD
-
Today on the blog we present results showing the operation of new dynamic circuits for quantum error correction, going beyond their static counterparts by using fewer couplers, removing correlated errors & utilizing a different type of quantum gate. Read more →https://goo.gle/4jDanqV
-
-
Announcing an advancement in our hybrid climate model, NeuralGCM, part of our Earth AI program to provide actionable insights about the physical earth to benefit society. We've trained the AI component directly on NASA satellite observations, leading to improved simulations of global precipitation, as published in Science Advances. This new approach allows NeuralGCM to more accurately predict: - Extreme Rainfall: Major improvements in capturing the intensity of the top 0.1% of rainfall events. - Long-Term Accuracy: A 40% average error reduction over land compared to leading climate models in multi-year runs. - Daily Patterns: Correctly reproducing the timing of peak daily precipitation. The open-source code and precipitation model checkpoint are now available for the community. Learn more: goo.gle/4qNBgL9
-
-
Building a long-lived logical qubit is essential for the next stage of quantum computing, and the next milestone on the Google Quantum AI roadmap. We put our Google Quantum AI researchers to the test for our "Hard Quantum Words" series to see how they break down this complex idea. Watch the video and tell us: Who explained it best?
-
What does quantum advantage really mean? In our new series, Hard Quantum Words, researchers from the Google Quantum AI team take on the challenge of explaining this foundational concept in plain language, showing how even the most complex ideas in quantum computing can be approached in a clear, understandable way. Watch how they break it down.
-
Advancing core scientific research, DeepSomatic & C2S-Scale join the AI-powered fight against cancer. Whether identifying genetic variants or generating novel hypotheses, our research helps scientists understand cancer & paves the way for brand-new therapies. Check out our review of 2025 for more ⬇️
In 2025, the magic cycle of research accelerated. Google Research teams delivered pioneering breakthroughs and brought our research to reality with impact on products, science and society. Read all about it in this blog from Yossi Matias, VP and Head of Google Research →https://goo.gle/48WNbyY
-