Shipping a Trillion Parameters With a Hub Bucket: Delta Weight Sync in TRL
Reachy Mini goes fully local
Algorithmic Monocultures in Hiring
Many employers screen job applicants with algorithms built by the same few algorithm vendors. We hypothesize that algorithmic monoculture leads to the same individuals and members of the same racial groups facing rejection. We acquire and analyze a novel dataset of 3 million applicants submitting 4 million applicati...
MUSE-Autoskill: Self-Evolving Agents via Skill Creation, Memory, Management, and Evaluation
Large language model (LLM) agents rely on reusable skills to solve complex tasks. However, existing skill creation approaches treat skills as isolated and static artifacts, limiting their reusability, reliability, and long-term improvement. We propose MUSE-Autoskill Agent (Memory-Utilizing Skill Evolution), a skill-...
LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding
Vision-language models (VLMs) commonly formulate visual grounding and detection as a coordinate-token generation problem, serializing each 2D box into multiple 1D tokens that are learned and decoded largely independently. This token-by-token decoding mismatches the coupled structure of box geometry and creates a pra...
Natural Language Query to Configuration for Retrieval Agents
Modern retrieval agents expose many configuration choices -- LLM, retriever, number of documents, number of hops, and synthesis strategy -- each shaping both answer quality and serving cost. Today, these pipelines are typically hand-tuned once per workload, leaving substantial per-query optimization untapped. We for...
GENESIS: Harnessing AI Agents for Autonomous 6G RAN Synthesis, Research, and Testing
Cellular research and development (R&D) is throttled by six structural processes that each consume months of manual engineering work per iteration: (i) synthesizing new features from standards or research papers into production code; (ii) conformance and interoperability testing; (iii) hardening against field anomal...
MobileMoE: Scaling On-Device Mixture of Experts
Mixture-of-Experts (MoE) has become the de facto architecture for hundred-billion-parameter language models, yet its advantages at sub-billion scales for on-device deployment remain largely unexplored. To close this gap, we present MobileMoE, a family of on-device MoE language models with sub-billion active paramete...
Alignment Tampering: How Reinforcement Learning from Human Feedback Is Exploited to Optimize Misaligned Biases
Reinforcement Learning from Human Feedback (RLHF) is the standard method to align Large Language Models (LLMs) with human preferences. In this work, we introduce alignment tampering, a potential vulnerability where the LLM undergoing alignment influences the preference dataset, causing RLHF to amplify undesired beha...
Guiding LLM Post-training Data Engineering with Model Internals from Sparse Autoencoders
Model internals encode rich information about how a large language model (LLM) processes its training data; however, post-training data engineering largely relies on external signals and ignores rich intrinsic signals lying in model internals. We propose SAERL, a data engineering framework for LLM reinforcement lear...
From Scores to Gibbs Correctors: Accelerating Uniform-Rate Discrete Diffusion Models
Discrete diffusion models have achieved strong empirical performance in text and other symbolic domains, but, especially for uniform-rate models, they often require many steps to generate a single sample. Existing acceleration methods either rely on training additional quantities or suffer from slow mixing. In this...
When Eyes Betray AI: Social Gaze Consistency as a Semantic Cue for AI-Generated Image Detection
Recent generative models have largely closed the gap on low-level artifacts - pixel fingerprints, frequency anomalies, upsampling traces - particularly in person-centric and partial-edit settings where the manipulated region is small and surrounded by photometrically authentic content. We introduce Social Gaze Consi...
Towards Controllable Image Generation through Representation-Conditioned Diffusion Models
Diffusion models have emerged as powerful tools for high-quality image generation and editing, but guiding these models to produce specific outputs remains a challenge. Conventional approaches rely on conditioning mechanisms, such as text prompts or semantic maps, which require extensively annotated datasets. In thi...
2-ASP(Q) programs with weak constraints: Complexity and efficient implementation
ASP(Q) extends Answer Set Programming (ASP) with Quantifiers over answer sets. In this paper we focus on the class of ASP(Q) programs with two quantifiers and weak constraints, denoted as 2-ASP(Q)^w. 2-ASP(Q)^w is a practically relevant fragment of ASP(Q) that is expressive enough to capture optimization problems up...
Probabilistic Smoothing with Ratio-Monotone Transforms for Global Optimization
Probabilistic smoothing is a standard tool for global optimization, but existing methods rely on Gaussian kernels and specific transforms, often resulting in strong hyperparameter sensitivity and limited robustness. We propose a general smoothing framework that combines flexible symmetric unimodal kernels with monot...
Greening AI Inference with Accuracy and Latency-aware User Incentives
The widespread use of AI services has raised concerns for its environmental sustainability, towards which recent studies have identified carbon emissions of AI inference as the major contributor. This paper introduces a framework for designing AI inference incentives based on the users' valuation for inference quali...
Normal Guidance is what Attention Needs
We consider training classifiers for 3D medical images using only one binary label for the entire volume rather than a label for each 2D slice. In such weakly supervised settings, can we learn accurate classifiers for slice-level predictions? Attention-based multiple instance learning (MIL) can produce an attention...
Harness, Scaffold, and the AI Agent Terms Worth Getting Right
Towards Speed-of-Light Text Generation with Nemotron-Labs Diffusion Language Models
Catch up on the Dialogues stage at Google I/O 2026.
Alphabet CEO Sundar Pichai in conversation on the I/O 2026 Dialogues stage
Specialization Beats Scale: A Strategic Variable Most AI Procurement Decisions Overlook
We’re launching the Google DeepMind Accelerator program in Asia Pacific to tackle environmental risks
We’re announcing new community investments in Missouri.
We’re helping build the state’s next-generation workforce and investing in energy programs.
100 things we announced at I/O 2026
Image with the words "Ready, Set, I/O" and a colorful Gemini logo