XOutlier
Download .zipAdd to Chrome

Google Research

@GoogleResearch119k followers

Impossible? Let’s see. From algorithms to neuroscience to AI, Google Research strives to progress science, advance society & improve billions of people’s lives.

Open on X
Top outlier
339×
Best organic post
Peak views
1.4M
Single post
Bookmarks
16k
Across 96 organic posts
Last scan
Oct 16
Community cache
Posts 96 posts
#DatePost↓ OutlierViewsLikesRepostsBookmarks
1Nov 7, '25Introducing Nested Learning: A new ML paradigm for continual learning that views models as nested optimization problems to enhance long context processing. Our proof-of-concept mod339×1.4M4.7k7852.9k
2Sep 12, '25Introducing VaultGemma, the largest open model trained from scratch with differential privacy. Read about our new research on scaling laws for differentially private language model82.4×348k1.5k230601
3Jun 3, '25Welcome to the new home for Google Research on X! To learn about the latest research from Google, read the most up-to-date Google Research Blog posts, or get details on our partici69.2×292k2.1k141311
4Oct 17, '25Scheduling virtual machines (VMs) is like a puzzle game with disappearing pieces. Here we present LAVA, an AI-based scheduling algorithm that continuously re-predicts & adapts 69×291k84897453
5Aug 7, '25Today on the blog we feature a new active learning method for curating high-quality data that reduces training data requirements for fine-tuning LLMs by orders of magnitude. Read a61.2×258k1.4k247939
6Sep 30, '25Today we describe how we leverage AlphaEvolve, a @GoogleDeepMind system for iteratively evolving code, to morph snippets of code towards better proof elements in complexity theory 60.9×257k1.2k194535
7Jul 9, '25Introducing new models for research & development of health applications: MedGemma 27B Multimodal, for complex multimodal & longitudinal EHR interpretation, and MedSigLIP, 36.9×156k996176352
8Oct 15, '25One of our new Gemma open models, Cell2Sentence-Scale, has identified a novel cancer therapy pathway that’s been validated experimentally in living cells. Developed w/ @GoogleDeepM32.1×136k1.4k206249
9Sep 19, '25TTD-DR is a Deep Researcher agent that models research report writing as a diffusion process, where a first draft is gradually polished into a high-quality final version. Read abou30.5×129k732135444
10Sep 16, '25Learn Your Way, now on Google Labs, is a research experiment that explores how GenAI can transform educational materials to create a more effective, engaging, learner-driven experi29.5×124k901167545
11Oct 16, '25For 10 years, Google has worked to accurately read the operating manual of all life on Earth — the genome. Our AI tools are now used by partners for real-world challenges from impr28.2×119k2k271396
12Oct 15, '25Introducing Coral NPU: A full-stack, open-source platform co-designed by Google Research and @GoogleDeepMind. It provides tools for developers to build the next generation of alway25.5×108k977166407
13Aug 6, '25Introducing DeepPolisher — in collaboration with UC Santa Cruz Genomics Institute, we developed this new open-source method to improve genome assembly accuracy. It reduces indel er24.7×104k793137182
14Jun 15, '25Congratulations to the authors of "Going Deeper with Convolutions", recipient of the 2025 Longuet-Higgins Prize, which celebrates a paper released 10 years ago that has had a signi22.3×94k1.2k124528
15Sep 9, '25Today, we’re launching AI Quests, a game-based learning experience for students to learn firsthand how AI can address real-world societal and scientific challenges, from forecastin18.2×77k1k160640
16Oct 22, '25Today we announce a significant step in quantum computing by introducing Quantum Echoes, a new quantum computational algorithm measuring out-of-time-order correlators. This verifia18×76k1.6k303290
17Aug 12, '25Today, Google Research & @GoogleDeepMind introduce g-AMIE, an extension of our diagnostic AI system based on #Gemini 2.0 Flash. It uses a guardrail that prohibits medical advic18×76k54694159
18Jul 28, '25Let your wearable data "speak" for itself! Introducing SensorLM, a family of sensor-language foundation models trained on ~60 million hours of data, enabling robust wearable data u17.7×75k1k137405
19Oct 23, '25Today, we are announcing the latest advancements to Google Earth AI — a platform designed to unlock a new level of planetary understanding. This includes new research on Geospatial16.6×70k65875158
20Jun 13, '25Improving tropical cyclone prediction can help communities better prepare & stay safe. Developed with @GoogleDeepMind, we're launching Weather Lab to share our AI weather predi16.4×69k947128206
21Jun 25, '25Neural embedding models have become a cornerstone of modern information retrieval. Today we introduce MUVERA, a state-of-the-art retrieval algorithm that reduces complex multi-vect14.4×61k38772206
22Jun 10, '25Floods impact millions of people annually. With AI we provide reliable 7-day flood forecasts, even in data-scarce regions. But the real impact comes from our partners — from govern13.8×58k1152122
23Nov 6, '25DS-STAR is a state-of-the-art data science agent designed to autonomously solve complex data science problems. It automates tasks from analysis to data wrangling across diverse dat12.6×53k790119424
24Oct 29, '25Introducing StreetReaderAI: A new and more accessible street-level imagery prototype using context-aware, real-time AI and accessible navigation controls. We're redefining immersiv12.1×51k59794197
25Oct 8, '25AlphaEarth Foundations is a new state-of-the-art model that revolutionizes global mapping and monitoring. This work already powers partners around the world to address challenges f11.6×49k1.1k142277
26Oct 7, '25Today's common voice search technologies are focused on the question, "What words were said?" What if we could answer a more powerful question: "What information is being sought?" 11.3×48k83496279
27Aug 18, '25Project Astra is a research prototype that explores the future capabilities of a universal AI assistant. At 4pm Aliya Rysbek will be at the #DLI2025 Google booth to demo Astra’s ca11.2×48k72585130
28Sep 11, '25Speculative Cascades is a new approach for improving LLM efficiency that combines the best features of both cascades (where a small LLM precedes a larger LLM) and speculative decod10×42k53879257
29Nov 17, '25Learn how our latest AI model, developed with @GoogleDeepMind, is helping us better predict the weather ⬇️8.6×36k3104540
30Jun 23, '25Error correction is a key component of tomorrow’s large-scale quantum computers. Today we’re excited to report the experimental demonstration of a color code system that provides a8.5×36k3055964
31Jul 17, '25Congratulations to Sergey Ioffe & Christian Szegedy, authors of "Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift", recipients of th8.0×34k1412427
32Jun 23, '25We’re announcing the 87 professors selected for the 2025 Google Research Scholar Program — join us in congratulating these exceptional recipients and learn more about their groundb7.8×33k93914
33Jul 10, '25Today on the blog we share our recent progress in developing graph foundation models that excel on interconnected relational tables and at the same time generalize to arbitrary set7.7×33k23438107
34Jun 5, '25Today we announce a novel generative AI model, dynamical-generative downscaling, that efficiently bridges the resolution gap between standard Earth system model outputs and what's 7.5×32k4437484
35Nov 7, '25Learn about Google’s approach to building AI to help improve learning outcomes for everyone in this new paper, co-authored by @ymatias, our VP and Head of Google Research ↓7.3×31k34361121
36Oct 16, '25DeepSomatic, an AI model developed with @ucscgenomics, identifies cancer cell genetic variants. In research with Children’s Mercy, it found 10 variants in pediatric leukemia cells 7.0×30k529115104
37Sep 23, '25We present a new approach to time-series forecasting that uses continued pre-training to teach a model to adapt to in-context examples at inference time, matching the performance o6.6×28k52276270
38Oct 27, '25Learn about the research behind our new personal health coach, built using Gemini models and thoughtfully developed with foundational research. Read more →6.2×26k3415694
39Oct 2, '25Ever had a perfect image in mind that a text-to-image model just couldn't capture? Our new reinforcement learning agent, PASTA, turns image generation into a collaborative conversa6.1×26k32958145
40Oct 1, '25Learn about our research prototype LLM-powered personal health agent that analyzes various data modalities, including data from wearable devices, to offer evidence-based health ins6.0×26k27544134
41Aug 6, '25Type 2 diabetes affects hundreds of millions globally. Our research explores a suite of ML models that show the potential of using wearable data & routine blood tests to effect5.7×24k41079123
42Jul 22, '25Introducing LSM-2, our newest foundation model for wearable sensor data. LSM-2 uses Adaptive & Inherited Masking, a novel self-supervised framework, to learn from incomplete da5.6×23k34555109
43Oct 20, '25Read all about how Google's Gemini model can be transformed into an expert astronomy assistant that classifies cosmic events and explains its reasoning in plain language, achieving5.4×23k4036974
44Jul 31, '25Our new regression language model approach enables LLMs to predict industrial system performance directly from raw text data. Learn how we applied it to Borg, Google’s large-scale 5.3×22k40374136
45Aug 1, '25Presenting MLE-STAR, a novel research focused ML engineering agent that integrates web search and targeted code block refinement that could help foster innovation and streamline ML5.1×22k31847117
46Sep 15, '25Excited to release TimesFM 2.5 on Hugging Face (soon available in BigQuery and Model Garden), significantly improving over TimesFM 2.0 in accuracy and maximum context length. Times5.0×21k2393485
47Oct 22, '25Stop by the #ICCV2025 Google both at 2:30pm HST for a behind the scenes look at how @GoogleDeepMind's AI research transformed the movie the Wizard of Oz to make an immersive 16K ex4.7×20k1352619
48Sep 18, '25Sensible Agent is a research prototype for proactive, unobtrusive AR agents that use real-time context (gaze, noise, hand availability) to offer seamless, socially-aware assistance4.4×19k1351924
49Sep 17, '25SLED is a decoding strategy that uses all of an LLM’s layers, instead of just the last one, to better align the output with the model’s intrinsic knowledge, enhancing model accurac4.3×18k1432152
50Nov 14, '25Next-gen telescopes will generate 10 million alerts a night. How can astronomers keep up? We prompted Gemini with 15 examples to classify cosmic events at 93% accuracy, & it ex4.1×17k1.3k93237
51Sep 25, '25Learn about user insights from a novel AI agent — “Wayfinding AI”, an early-stage research prototype — that helps find the way to better health information through proactive conver3.9×17k2354193
52Aug 14, '25Generating large-scale differentially private (DP) synthetic data is challenging due to the fundamental privacy–computation–utility trade-off. Here we describe CTCL, a novel, more 3.7×16k2824973
53Jul 18, '25Using aggregated measurements from a global network of Android smartphones, we developed a system that detects earthquakes, delivers early warnings to users, and builds user trust 3.6×15k3025757
54Jun 5, '25At yesterday's The Beyond Series: Research@Tokyo 🚀 with researchers, policymakers, & industry leaders, we explored cutting-edge research in health, social innovation, & hu3.6×15k61143
55Oct 23, '25Check out this new Featured Notebook to dive into our genomics research and learn more about how studying the genome can improve healthcare ↓3.5×15k1752629
56Jul 15, '25Today at the #ICML2025 Google booth at 11:30am, researchers will present Project Astra, a research prototype exploring breakthrough capabilities for Google products — on the way to3.2×13k2784250
57Oct 9, '25Today on the blog learn about XR Blocks, an open source framework to author #XR + #AI applications on #AndroidXR with WebXR → You can check out over 20 live demos at !3.2×13k601719
58Oct 16, '25What if we could use AI to identify genetic variants in cancer cells? DeepSomatic uses ML to find key DNA variants with higher accuracy, aiming to create a better understanding of 3.0×13k2604258
59Nov 14, '25What if we could identify communities most vulnerable to a hurricane before it makes landfall? Using our state-of-the-art Earth AI models of imagery, population, and environment da2.9×12k2153642
60Sep 24, '25Ensuring generalization of LLMs in response to distribution shifts is especially important for medical and health-related models. Here we describe AfriMed-QA, an open-source benchm2.7×12k1371830
61Nov 13, '25Today on the blog, learn about new theoretical work from @GoogleQuantumAI showing that large scale quantum computers could solve certain optimization problems that are intractable 2.6×11k2283734
62Aug 27, '25What if we could predict tropical cyclones with more accuracy & up to 15 days before? Our new experimental AI model, Weather Lab, developed in partnership with @GoogleDeepmind, mak2.5×11k2433543
63Nov 12, '25Here’s the release of JAX-Privacy 1.0, a library for #DifferentiallyPrivate #ML on the high-performance JAX computing library, making it easier for researchers & developers to 2.5×11k1743232
64Aug 20, '25At 4pm today, Sekou Remy & Mercy Asiedu will be at the #DLI2025 Google booth to demonstrate applications of MedGemma, a suite of our most capable open models for health AI deve2.4×10k2263756
65Oct 16, '25Google is proud to be an Ultimate Sponsor of #ICCV2025, happening Oct 19–23 in Honolulu! Google Research has a strong presence at @ICCVConference — learn about our work, including 2.2×9.2k1171410
66Nov 14, '25What if we could anticipate forest loss before it happens, rather than just counting it after the fact? Introducing ForestCast: a new model from @GoogleDeepMind and @GoogleResearch2.1×8.9k42342114
67Oct 31, '25At Research@ in Mountain View last week, we shared some of our latest announcements, from understanding earth to advancing genomics & quantum computing. Read all about the powerful2.0×8.7k1351529
68Jul 30, '25Announcing AlphaEarth Foundations, part of our Earth AI collection of geospatial models that tackle critical planetary needs: ☁️ Weather 🌊 Floods 🔥 Wildfires 🏙️ Urban planning �2.0×8.7k2453825
69Oct 9, '25Excited to share that FireSat is one of TIME's Best Inventions of 2025! A collaboration between the Earth Fire Alliance, Muon Space & Google Research, it’s an AI-powered satell2.0×8.3k1111719
70Aug 27, '25Check out today’s blog post, where we highlight two studies that explore how Google’s AI models can be used as helpful tools in medical learning environments, reimagining the futur2.0×8.3k1352026
71Oct 30, '25To ensure strong transparency in user data handling and analysis, today on the blog we detail how confidential federated analytics technology is leveraged to understand on-device g2.0×8.3k1392323
72Sep 29, '25What if we could predict tropical cyclones with more accuracy up to 15 days in advance? Our new experimental AI model, Weather Lab, developed with @GoogleDeepmind, makes prediction1.9×8.2k2663465
73Nov 13, '25Together with @GoogleDeepMind, we present Natural Forests of the World 2020, a new map & dataset that distinguishes natural forests from other tree cover with 92.2% accuracy, creat1.9×7.9k1693228
74Nov 5, '25Together with @GoogleDeepMind we’re releasing ForestCast, a new benchmark & paper on forecasting deforestation risk with deep learning. This shifts from counting forest losses 1.9×7.9k1393323
75Oct 22, '25Google recently released Nano Banana, the latest Gemini native multimodal generation model. Check out a demo today at 3:30pm HST at the #ICCV2025 Google booth, highlighting some of1.8×7.8k1062213
76Nov 5, '25Researchers at Google are using AI to map Earth’s biodiversity. Learn how our complex deep learning architecture, known as Graph Neural Networks, is upleveling species distribution1.8×7.7k3004071
77Oct 21, '25Are video models the path to general visual intelligence? Check out the #ICCV2025 Google booth today at 3pm to see how Veo 3 solves tasks for which it wasn't trained.1.7×7.1k5476
79Nov 14, '25StreetReaderAI is making street-level navigation more accessible for blind and low-vision users. Our new prototype uses context-aware, real-time Gemini models to convert Street Vie1.6×6.9k53045123
80Aug 20, '25We present novel algorithms for differentially private partition selection, which is an approach to maximize the number of unique items selected from a union of sets of data, while1.6×6.8k1472428
81Oct 8, '25A key challenge for LLMs is factuality. SLED is a new decoding strategy from Google Research that helps to improve accuracy by using all of a model's layers to align its output wit1.5×6.4k19724107
82Jul 23, '25First wildfire images from FireSat are here! FireSat spotted an Oregon fire undetected by existing systems, showing its ability to catch small fires in their earliest stages — an i1.5×6.3k1492420
83Aug 21, '25Discover how we distill massive models into mobile-friendly experiences for YouTube Shorts! The model preserves identity with Pivotal Tuning Inversion & optimizes with MediaPip1.5×6.2k1191624
84Sep 29, '25Floods impact millions annually, and we're using AI to provide reliable 7-day forecasts, even in data-scarce regions. Our partners — from governments to aid orgs — are the ones who1.4×6k2913475
85Jul 2, '25Today we describe SpeechCompass, an exploration that uses multi-microphone localization to enhance mobile captioning with speaker diarization (i.e., separating speakers in an #Auto1.4×5.8k821323
86Aug 26, '25Evaluation of language models in complex domains (such as health) can be expensive and labor intensive. We present a new adaptive and precise rubric methodology that saves time and1.4×5.8k1142630
88Oct 22, '25We're excited to announce that Android XR is now available on Samsung's new Galaxy XR! And, if you're attending #ICCV2025 you can experience it first-hand at the Google booth on Th1.2×5.2k82157
89Jun 4, '25Today on the blog we propose Action-Based Contrastive Self-Training, a data-efficient #ReinforcementLearning tuning approach for improving multi-turn conversation modeling in mixed1.2×5.2k81724
91Aug 20, '25Update: Our Q&A session about Gemini & post AGI research at Google DeepMind is rescheduled for tomorrow morning, Wed. at 11:30am. See you then!1.2×5k6247
92Jul 17, '25Today on the Research Blog, we describe how transfer learning can enable contactless heart rate (HR) monitoring via ultra-wideband radar, paving the way for possible future develop1.1×4.8k1192031
93Nov 13, '25Ever wonder how scientists know what’s in your genome? Explore the latest @NotebookLM Featured Notebooks grounded in Google Research. Learn about DeepVariant, genetic diagnosis, se1.1×4.7k75919
94Oct 21, '25Experience Android XR firsthand at the #ICCV2025 Google booth today at 12pm! See the advanced Perception stack for seamless immersive experiences, translating vision research into 1.1×4.7k76178
95Oct 23, '25Missed us on Tuesday? Here's another chance to experience Android XR firsthand at the #ICCV2025 Google booth today at 11:30am! See the advanced Perception stack for seamless immers1.0×4.2k56107
96Oct 20, '25To address the need for synthetic versions of rich, structured image-based datasets, we introduce a hierarchical method for generating differentially private synthetic photo albums0.9×3.9k5348
97Jul 22, '25Congratulations to the authors of "Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction", which won an #ICML2025 Outstanding Paper Aw0.9×3.8k501325
98Sep 11, '25Introducing NucleoBench, the first standardized benchmark for nucleic acid design algorithms. We also present a new algorithm, AdaBeam, that outperforms existing methods on 11 of 10.8×3.3k50312
99Aug 18, '25Attending #DLI2025? Stop by the Google booth to learn about some of our groundbreaking research by attending a demo or chatting with researchers. You can find the schedule of our @0.8×3.2k5441

Scan another creator

@
@karpathy@levelsio@paulg@naval@marclou@illyism

Run this on any account.

XOutlier builds this table for every public profile you open on x.com, with live data and every post, not just the top 50.

Add to Chrome Download .zip