A developer built PA-Trace, an on-device agentic workflow using Google's MedGemma model to automate prior-authorization documentation for imaging orders. The system extracts evidence from clinic notes…
Google Research is using AI to accelerate scientific discovery across genomics, neuroscience, epidemiology and climate science, according to Lizzie Dorfman, who leads science AI at the company. Dorfma…
Google Research introduced TabFM, a foundation model for tabular data that performs zero-shot classification and regression without dataset-specific training. The model uses a hybrid-attention archite…
Google Research introduced TabFM, a zero-shot foundation model for tabular data classification and regression, integrated into BigQuery ML. The model uses in-context learning to eliminate manual train…
A developer explains the inner workings of the ReAct (Reason + Act) pattern for AI agents, distinguishing between the message-level and state-level perspectives. The core of ReAct is not the prompt fo…
Sean Xuefeng Du will present a new hallucination-detection method, Answer-agreement Representation Shaping (ARS), at ICML 2026 in Seoul on July 9. The work, developed at Nanyang Technological Universi…
Google Research head Yossi Matias outlined the company's vision for AI-driven science, emphasizing tools like Gemini for Science and AI Co-Scientist that aim to amplify human ingenuity rather than rep…
Synaptics and Google Research debuted the Coralboard edge AI development platform at Google I/O 2026, featuring the Astra SL2610/SL2619 SoC family and integrated NPUs. The platform supports on-device …
Shrijith Venkatramana, building git-lrc, explains the evolution of LLM evaluation metrics from BLEU to BLEURT and COMET. BLEU, introduced in 2002, measures n-gram overlap and correlates with human jud…
Google Cloud and Google Research introduced linear elastic caching, a new approach that dynamically adjusts cache size using machine learning to minimize total cost of ownership. The method frames pag…
Researchers at Google Research found that allowing large language models to generate reasoning traces improves recall of simple factual knowledge, even when no complex reasoning is needed. The phenome…
Google Research developed MoGen, an AI model that generates synthetic neuron geometries to improve brain mapping. The model reduced reconstruction errors by 4.4%, potentially saving 157 person-years o…
Google Research published a study in Nature showing its Articulate Medical Intelligence Explorer (AMIE) extended from diagnosis to longitudinal disease management, matching clinicians in overall manag…
Google's AMIE diagnostic AI system matched or outperformed primary care physicians on 30 of 32 metrics in a randomized study published in Nature, with patient-actors rating it superior on 25 of 26 mea…
Google Research released a vectorized dataset of fine-scale ecological features across England, including hedgerows and copses, using a deep learning framework to reveal features invisible to standard…
Google Research released Gemini-SQL2 on June 12, a text-to-SQL system built on Gemini 3.1 Pro that achieved 80.04% execution accuracy on the BIRD benchmark, the first model to surpass 80%. The system …
Google Research's Gemini-SQL2, built on Gemini 3.1 Pro, achieved 80.04 percent accuracy on the BIRD text-to-SQL benchmark, surpassing OpenAI and Anthropic. The technology aims to enhance natural langu…
Google Research scientists Rory Sayres and Yun Liu published findings in JAMA Dermatology showing that AI tools can help laypeople better identify skin conditions and determine appropriate next steps.…
Researchers at Google Research introduced Regularized f-Divergence Kernel Tests, a new framework for auditing machine unlearning that improves sensitivity and accuracy in verifying whether AI models h…
Google Research and Google Cloud introduced a new agentic retrieval-augmented generation (RAG) framework hosted on the Gemini Enterprise Agent Platform, designed to break down complex enterprise queri…