AI research papers.

Vision-language-action (VLA) models transfer semantic knowledge acquired through vision-language pretraining to robot control, achieving strong performance across various benchmarks.

As coding agents become increasingly capable, robust evaluation requires a broader and more diverse supply of verifiable software engineering (SWE) tasks.

While generalist vision-language-action (VLA) models have enabled diverse open-world manipulation, real-world robotic assistance requires two critical capabilities they currently lack: personalization and reliable execut…

Unified multimodal models (UMMs) receive image-generation supervision that exposes them to visual information not fully documented in text.

Large language model agents solve tasks through long sequences of tool calls, and the prevailing design has a single strong model generate every step, although many steps are routine enough for a much smaller model.

Transcribing expressive piano performances into scores requires interpreting individual note timings within an underlying metrical structure.

ARTIFICIAL INTELLIGENCE (AI) has demonstrated its potential across a wide range of humanitarian tasks, offering scalable and data-driven solutions to some of the world's most pressing challenges.

Reusing the attention KV cache across requests is a standard way to cut prefill cost in long-context LLM serving, and production systems increasingly back the per-engine cache with a remote tier shared by every engine.

Large language models (LLMs) deployed as conversational agents are increasingly expected to follow procedural logic.

Policy-shift distillation reuses the difference between a teacher's log probabilities before and after reinforcement learning (RL) as supervision on student-generated trajectories.

Data agents use code to process heterogeneous files and derive analytical insights.

Can AI agents organize scientific investigations that lead to well-supported conclusions?

We study how vision-language models (VLMs) can learn to think and actively explore in 3D space.

Large language model (LLM) agents have demonstrated strong capabilities in complex tasks, including code generation, tool use, and multi-hop question answering.

As Vision-Language Models (VLMs) advance, they are increasingly expected to perform spatiotemporal reasoning over motion, actions, and view changes beyond what the input images show.

Self-evolving language models train on tasks they generate, so every round must decide which updated model the next round inherits.

Streaming vision-language models (VLMs) must continuously integrate information from the past while remaining responsive to current observations and user queries.

Personalization of Large Language Models (LLMs) is increasingly important for user-centric applications, yet prior methods mainly adapt dense LLMs, whose polysemantic neurons, each encoding multiple unrelated concepts, h…

Interactive agents repeatedly choose among a finite set of actions, labels, or ordinal outcomes.

The vocabulary of a language model is its native semantics: the embedding and the unembedding fix the meaning of a state at the two ends of the residual stream.

Most existing agentic reinforcement learning methods assume single-session trajectories, where interaction proceeds within one continuously accumulated context.

Recent advances in Omni Language Models (OLMs) provide a unified solution for audiovisual video captioning.

Converting verifier outcomes into useful policy updates and partial progress into complete task resolution is crucial for terminal agents.

Language model agents increasingly automate post-training, but when given the full training stack they converge on supervised fine-tuning and spend most effort revising training data.

Computer use has reached a turning point: the focus is expanding from GUI-only interaction to systems that complete tasks on users' computers through visual and programmatic interfaces.

Reinforcement learning (RL) can elicit specialized visual capabilities in vision-language models (VLMs), yet jointly optimizing heterogeneous objectives often introduces cross-capability gradient conflicts.

Real desktop automation is neither purely visual nor purely programmatic.

Multimodal deep search is moving from short queries to long-horizon investigations that continually acquire evidence.

Reinforcement learning (RL) enables language models to improve text-to-image (T2I) generation by rewriting prompts for a frozen image generator.

Reasoning over long document contexts remains a challenge for large language models (LLMs).

Spatial intelligence lets an agent perceive and act in the physical world, and many of its judgments must be made at interaction rates: which spatial relation holds, how far an object is, or where to place it.

Computer-use agents increasingly tackle long-horizon tasks, where running a frontier model at every step is expensive and a small open-weight model alone often fails.

Reinforcement learning can align visual generators with reward signals beyond their pretraining objectives, but existing score-function methods typically introduce stochasticity by perturbing the reverse denoising trajec…

Multi-object tracking in sports broadcasts is hard for two reasons: athletes move fast and change direction abruptly, and the camera pans and zooms at the same time.

Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation.

We placed four Gemini 4 Argon agents in a sandboxed Minecraft survival world and gave them one goal: defeat the Ender Dragon.

As joint audio-video generation models become more capable, they can generate audio-visual scenes exhibiting three forms of complexity: multi-event composition, multi-layer audio mixture, and asymmetric cross-modal relat…

The capabilities of an agent swarm depend on both its individual members and how they coordinate.

One of the first steps of a research pipeline is generating ideas.

Image editing models produce impressive results for individual edits, but their quality often deteriorates over multiple editing turns.

We introduce 3DCodeVerse, a research ecosystem that enables coding agents to create editable and verifiable 3D content at scale.

Recent 3D world models increasingly use features from geometry foundation models (GFMs) as predictive world states, benefiting from their rich geometric priors.

Large language models (LLMs) are expanding from assisting scientific research and engineering to autonomously developing and improving AI systems, advancing AI for AI (AI4AI) as a pathway toward recursive self-improvemen…

Few-step autoregressive video generators enable streaming generation, yet they run every token through the final denoising step, regardless of how much refinement it needs.

Existing agent benchmarks typically evaluate agents on well-specified tasks defined by issue reports, specifications, or task tickets.

Single image 3D generation aims to dress a given 3D mesh with relightable materials while faithfully preserving the visual details of a reference image.

Generative visual verifiers play a crucial role in advancing multimodal intelligence.

Time series foundation models are becoming increasingly popular for general-purpose forecasting.

Generative simulation can automate robot environment construction, but physically valid environments are not necessarily useful for training.

We introduce Gen4P, a unified framework for recovering 4D geometry and motion from monocular video.