Impossible? Let’s see. From algorithms to neuroscience to AI, Google Research strives to progress science, advance society & improve billions of people’s lives.
Joined May 2017
- Tweets620
- Following17
- Followers122K
- Likes188
Good medicine isn't just about giving the right answers — it's about asking the right questions.
See how @GoogleResearch collaborates with clinicians to train models that understand patient context: youtu.be/4XzcdKvF60A
Today on the blog we introduce an investigational ML framework that translates passive wearable signals into a modifiable coachable health age, uses an action-conditioned world model to simulate personalized 12-week "what-if" habit trajectories, & expands to routine blood biomarkers to track early cardiovascular-kidney-metabolic (CKM) health.
Learn more: goo.gle/4zFlqqF
ALT Overview of our end-to-end framework, from estimating coachable health age and predicting long-term disease risk to simulating 12-week "what-if" habit trajectories for personalized coaching.
Huge congratulations to our own Juliet Rothenberg on being named one of @TIME's 100 climate leaders of 2026! 🌍 The #TIME100 Climate list honors the leaders driving climate action in business. More here: goo.gle/4sPE8JI
Google Research retweeted
Evaluating generative technology for health functions takes more than multiple-choice tests.
Our new whitepaper details our framework combining benchmarks, automated assessments, and expert human review to assess accuracy, safety, and usefulness. goo.gle/3UxGBfg
Fewer than 100 of the world’s 7,000+ languages are well represented online. Our interactive web platform, Language Explorer, helps bridge this gap with rich data & visualizations.
Learn more & read about a new exhibition at the Museum of Communication in Berlin, featuring Language Explorer & more: goo.gle/4zwJ6xz
We’re excited to officially expand our global research footprint with the launch of our new Google Research hub in Singapore.
As @ymatias, VP, Google & GM of Google Research, shares below, this new hub focuses on three core areas: advancing foundational AI efficiency alongside our global teams (a critical priority for generative AI), cultivating local scientific talent through close partnerships with the National University of Singapore (NUS) and Nanyang Technological University (NTU), and scaling existing Google Research health initiatives by bringing open models like MedGemma to frontline care.
Read more in our blogs:
Advancing research and partnering with universities in Singapore: goo.gle/4fdW8I1
AI for nation’s healthcare system: goo.gle/4A72IIy
I spent a few days in Singapore energized by the incredible teams and the ecosystem.
Excited to share that we’re expanding our footprint with the launch of our new @GoogleResearch hub in Singapore.
Starting with a team that will work on advancing frontier AI efficiency alongside our global teams.
We are also partnering with local universities like @NUSingapore and @NTUsg to nurture the next generation of scientific leaders.
🧵👇
Most of what shapes human health happens outside the doctor's office: in daily sleep, diet, and continuous physiological signals.
What happens when AI helps connect those dots into a multi-year "health weather forecast"?
A breakdown of the latest @GoogleResearch podcast👇
Key milestones discussed:
• Open-source impact: How MedGemma powers local community health workers across the globe
• Imaging workflows: Why reducing mammography read times from days to minutes can potentially help save lives
• Frontier research: Inside AMIE, Google's multimodal diagnostic AI
1
24
3,656
A central takeaway: AI isn't replacing physicians: it's helping automate administrative burden so doctors can focus on what matters most: holistic reasoning, critical judgment, and patient empathy.
Watch the full conversation: youtu.be/4XzcdKvF60A
Today on the blog, we present a new system, published in Nature, that leverages standard consumer smartwatch sensors to passively detect signs of high blood pressure — validated against the clinical gold standard of 24-hour ambulatory bp monitoring. More →goo.gle/4zw2ISD
ALT The AI system uses photoplethysmography (PPG) and accelerometer (ACC) signals recorded using a consumer smartwatch to create waveform embeddings. The waveform embeddings are aggregated over time (days) into a person-level embedding, which, along with a user’s height, are used to predict their blood pressure status.
Introducing Agile Deliberation, a framework that turns vision-language models into collaborative partners for defining gray-area categories — from content moderation to curating art — while reducing user fatigue & improving classifier accuracy. Read more: goo.gle/4BHNAmr
In the AI era, the most critical skill isn’t just knowing how to code — it’s knowing which questions to ask.
@sineadbovell sat down with @ymatias (VP, Google & GM of @GoogleResearch) to discuss the rise of ambient intelligence, cross-disciplinary builders, and how researchers are evolving into "organizations of one."
Watch the full conversation:
youtu.be/FlOufV3hZFo
The most powerful technology is the kind you stop noticing.
Futurist @SineadBovell sat down with VP, Google & GM, Google Research @ymatias to unpack why the future of AI is ambient intelligence:
→ What felt like science fiction yesterday becomes today's invisible baseline
→ AI fades into the background as reliability reaches critical mass
→ Human ambition moves upstream to solve larger, more complex challenges
Watch the cutdown above and catch the full episode of I've Got Questions on YouTube: youtu.be/FlOufV3hZFo
What does hiring look like when everyone has access to an AI "virtual lab"?
Futurist @SineadBovell sat down with VP, Google & GM, Google Research @ymatias to break down why the next generation of builders must think like an "organization of one":
→ Asking the right questions
→ Connecting dots across disciplines
→ Applying ethical judgment
→ Adapting as intelligence becomes ambient
Watch the full conversation on I've Got Questions:
youtube.com/watch?v=FlOufV3h…
AI acts as an amplifier of human ingenuity.
A look inside Research@ 2026 in Mountain View — from open tools like Deep Review to foundation models across data science and planetary health, see how we're putting breakthrough capabilities in the hands of experts 👇
🔗 goo.gle/4hU5Srt
Read below on how our teams compressed months of low-level RDMA networking and custom TPU kernel development into hours ⬇️
Today, we shared how Google’s Graph Mining and Graph Learning teams used Gemini 4 Argon with Teamwork agent in @GoogleAntigravity to solve a different class of internal engineering bottleneck: compressing months of low-level RDMA networking and custom TPU kernel development into hours. Specifically, improvements include accelerated graph algorithms with FarArray (up to 180x faster, 22x less RAM) and custom TPU kernels for Graph Neural Networks that run 8–15x faster than native JAX with massive HBM savings.
For me, this update is a full-circle experience: Graph Mining contributed deeply to Gemini Data via deduping and data curation over time, and now Gemini is giving back by scaling and improving Graph Mining library!
This is also a close collaboration amongst our Graph Mining and Teamwork teams, together with the Antigravity and Gemini teams. Congratulations!
Read the full blog: antigravity.google/blog/gemi…
Today on the blog, learn about our new instance of a “sleep” phase in continual learning, which enables LLMs to continuously integrate new facts and enhance reasoning while mitigating catastrophic forgetting or static knowledge cutoffs.
More →goo.gle/4iICkyD
ALT Overview of our sleep framework: Input streams from the active phase are stored in short-term memory before entering the sleep cycle, which bifurcates into Consolidation (transferring fragile short-term tokens into expanded long-term parametric memory using low-rank structures and MoE) and Dreaming (autonomous self-improvement via synthetic routing with random experts and RL refinement).
Verena Rieser is heading to the @emnlpmeeting Industry Panel today at 11:00am to continue the discussion on agent reliability, trust, and deployment. Join us afterward at the #EMNLP2026 Google booth (#D5) to connect with our team. @GoogleDeepMind
How can positive alignment anchor autonomous AI agents to core values rather than brittle rubrics? Join @GoogleDeepMind researcher Verena Rieser today at 9:00am for the @emnlpmeeting Industry Keynote, and visit the #EMNLP2026 Google booth (D5) afterward to connect with our team.
