AI research papers.

Even a carefully proofread paper can leave small gaps in its proofs or its stated assumptions.

Mixture-of-experts (MoE) transformers, the standard backbone for modern large-scale deep learning, employ conditional computation: only a sparse subset of model weights is used in each forward pass.

Speculative decoding is a technique that accelerates the inference speed of autoregressive large language models (LLMs) by sampling from a faster, lower quality draft model.

Phenomena responsible for emerging technologies originate from novel molecular states at buried interfaces.

We introduce TWIG (Tree-structured Weights with Inherited Gates), a representation in which every weight of a neural network is the end of a path of small discrete decisions: a sign, a shift by a power of two, a zero gat…

The rapid growth of microdramas has increased demand for high-quality script production.

Large language model agents have enabled software development through natural-language interaction, but extending this paradigm to end-to-end game development remains challenging.

See how OpenAI, PerturbAI, and Scripps Research used distributed AI agents to uncover promising research leads in data from millions of brain cells.

arXiv:2605.21187v2 Announce Type: replace-cross Abstract: As distributed model training scales to span hundreds of thousands of GPUs, scale-out networks face unprecedented performance and efficiency demands.

This paper shows that JEPA-style latent world models (LWMs) can learn across diverse domains and data sources by adopting the language-conditioned interface of video generation models.

Online scene change detection requires integrating incoming observations to determine which regions have changed relative to a reference scene.

Text-to-image (T2I) platforms typically return several images per prompt so users can explore different visual interpretations and choose one that matches their intent.

Frontier black-box Large Language Model (LLM) agents are proficient at long-horizon reasoning and precise action generation.

Spatial reasoning requires more than access to geometric tools: a reasoner must decide when direct observation is sufficient and when to reconstruct, measure, or change reference frames.

Autonomous agents powered by foundation models can reason, plan, and use tools, but their fixed weights and context limits hinder sustained learning from experience over long-term deployment.

Neural speech codecs provide discrete tokens for speech generation.

Full 3D attention is costly in video Diffusion Transformers.

Agent harnesses have become a central component of modern agentic systems, shaping how models use tools, manage context, and coordinate long-horizon interaction.

Codes of professional ethics guide and restrict specialists, who increasingly work with assistants built on large language models.

Fine-grained multimodal intelligence requires models to both understand a designated region and localize it from language.

Recursive self-improvement (RSI) describes a system that improves not only its behavior but also the process by which it improves.

Using an LLM to generate training data for another is among the most general ways to train one, and it remains common practice.

Multimodal large language models (MLLMs) show promising reasoning capabilities, but reliable physical interaction requires planning that accounts for when actions are executable and how they change the world.

Most contemporary artificial intelligence (AI) systems are released with frozen parameters and improve through external updates, whereas an ideal system would keep improving from its own experience, a capacity we call re…

Diffusion-based video generation models (VGMs) can produce visually plausible videos, yet their camera ego-motion often drifts over long horizons.

A benchmark is not merely a collection of test questions; it naturally induces an empirical measure over difficulty.

Can LLM agents make contributions to frontier computer science R&D?

As LLMs expand their language coverage, evaluating how well they understand and represent diverse cultures becomes increasingly important.

Autoregressive video diffusion models generate videos chunk by chunk, yet producing plausible continuations does not guarantee that they capture how scenes evolve.

AI-generated text detectors are increasingly deployed in consequential settings---peer review, education, admissions, journalism, and content moderation---where their outputs directly affect individuals.

Robots must act under physical and social constraints while adapting to changes in their environment and human needs.

Uniform diffusion language models (UDLMs) generate text by iteratively denoising sequences of ordinary vocabulary tokens, any of which may be corrupted and subsequently revised.

Perimeter triage faces a latency/semantics trade-off: signature engines are fast and format-safe but blind to unseen attacks, while generative models are semantic but slow and produce unparseable output.

World models turn visual representations into predictions and actions.

Reinforcement learning (RL) is an effective way to strengthen the long-context reasoning of large language models (LLMs), but it needs training data that are long, diverse, challenging, and verifiable.

Recent world action models provide visual foresight and future prediction for robot manipulation, but lack explicit mechanisms for compositional skill reuse.

World Action Models (WAMs) have emerged as a promising approach to learning generalist robot policies using rich spatiotemporal priors from pre-trained video diffusion models.

We introduce Foundation4D, a unified feed-forward model for geometry estimation, 3D detection, semantic occupancy, and Gaussian reconstruction from multi-camera driving video.

Modern LLM agents can retrieve information from accessible documents, and then disclose a sensitive conclusion assembled from them.

Recent advances in search agents have enabled large language models to tackle complex tasks by retrieving and synthesizing external information.

Reproducing a published 3DGS paper takes weeks of expert engineering before any extension can be proposed.

Lyapunov functions and barrier certificates give state-space guarantees of stability and safety, but synthesizing and verifying them---especially for nonlinear, high-dimensional systems---remains an open challenge.

Large-scale AI training increasingly relies on massive video datasets with random temporal sampling patterns.

Lab Director Chris White has worked on research challenges with real-world implications—from new approaches to wartime data analysis to tools for combating human trafficking.

Mixture-of-experts (MoE) models often exceed a single accelerator's memory capacity.

A fine-tuned Vision-Language-Action (VLA) policy can easily place the alphabet soup in the basket on command, just as it saw during fine-tuning.

In a hierarchical multi-agent system, a Leader assigns tasks and subagents carry them out.

Language models can make personalized recommendations from rich item content, user history, and feedback.

In long-term interactions with large language model (LLM) agents, users reveal habits through their choices and actions.