AI research papers.

A vision--language--action (VLA) policy can complete complex tasks while ignoring the evidence that should determine its actions.

Numerical Linear Algebra (NLA) has consistently played a vital role in advancing science by providing tools to solve fundamental problems encountered in scientific and engineering applications.

Recently proposed Drifting Models shift iterative distribution refinement from inference to training, enabling effective one-step generation.

Sparse-view novel view synthesis is a central problem in 3D content creation, but diffusion-based approaches remain limited by iterative denoising, making multi-view generation expensive at inference time.

We present DistScene, a framework for single-image compositional 3D scene generation by jointly modeling the environment and individual objects.

arXiv:2610.00902v1 Announce Type: cross Abstract: One way to make AI systems safe is to shape what the system is: its objective and dispositions.

Intent-based Open RAN needs an interpreter that turns intents into A1 policies within the loop of the RAN intelligent controller (RIC).

Cross-lingual contrastive learning has been a core component of multilingual encoder training, but the ability to explicitly align representations is not possible in decoder-only LLMs because of varying multilingual toke…

Full-duplex voice agents make many small, closed decisions, which current systems answer by slow autoregressive decoding.

Terminal agents are increasingly deployed beyond software engineering in science and other specialized domains.

Reinforcement learning for long-horizon agents relies on purely retrospective training signals: credit is assigned only after observing environmental consequences, leaving the agent's belief at action time invisible to t…

Large language models (LLMs) increasingly serve as general-purpose interfaces to factual knowledge, but their parameters do not automatically reflect information that changes after pretraining.

Many useful language-model tasks cannot be evaluated by exact outcome verification.

Transferring robotic skills from simulation to reality requires task knowledge that remains usable across differences in perception, dynamics, and embodiment.

Rubric-based reinforcement learning extends reward-driven optimization to open-ended tasks by assigning partial credit to individual response requirements.

Successful trajectories on difficult tasks provide valuable supervision for model improvement, but specialized harnesses introduce interventions that may be unavailable during deployment.

Vision-Language-Action (VLA) foundation models have scaled rapidly to enhance manipulation performance and generalizability, but this scaling incurs high computational costs that render real-world deployment increasingly…

World action models jointly learn to forecast world dynamics and predict robot actions, such that the learned internal world dynamics guide accurate actions.

Language model agents are deployed with a harness, the software around the model that manages its context, tools, and feedback.

Recent text-to-image models have made substantial progress in photorealism, aesthetics, and text-image alignment.

Recent text-to-image generation models have achieved remarkable visual quality, but improving them through post-training remains challenging because no single reward signal captures the full range of human preference.

Long-form thinking traces can substantially improve the multi-step reasoning performance of large language models (LLMs), but they introduce high inference-time overhead, with latency dominated by sequential decoding.

Modern autoregressive (AR) video diffusion models excel at short-horizon video generation, yet generating long videos remains challenging due to drifting, where colors and textures shift, and motion dynamics decay.

Existing streaming vision-language models (VLMs) continuously perceive and reason over visual streams, but their computational pathways remain fixed throughout inference.

On-policy distillation (OPD) has become an important approach to language model post-training.

As LLM agents decide on users' behalf which product to buy, which hotel to book, or which paper to cite, a preference for items from certain sources (the sites or services they come from) shapes what users receive and wh…

Large language models (LLMs) are increasingly deployed in public health, finance, and governance, requiring both accuracy and societal value alignment.

Generative models of human-object interaction are bounded by the data that exists: everyday activities involve several objects, but most captured datasets record one at a time, as multi-object capture is combinatorially…

We reconstruct high-fidelity 3D scene meshes from posed multi-view images without per-scene optimization, across scales ranging from single objects to large outdoor scenes.

We understand little about how capabilities acquired in one language carry over to another, or what governs this transfer: evaluations rely on incomparable, saturation-prone datasets and rarely examine its determinants j…

World action models integrate future visual dynamics with robot action prediction, but their scalability remains limited by the need for action-annotated robot trajectories.

Chain-of-thought (CoT) reasoning allows humans to inspect how large language models reach their answers, and oversee model behaviour.

Streaming video generation has benefited from distribution matching distillation (DMD), which matches the joint distribution of video frames to a video teacher's approximation of the real video distribution.

We introduce HyperBrowseComp, a multilingual and multimodal browsing benchmark comprising 423 manually authored and human-validated questions across 13 languages, written by native or highly proficient speakers.

Learning general-purpose robot policies requires large-scale real-world interaction data, yet collecting robot demonstrations remains expensive and difficult to scale.

Image-text alignment is a core problem in computer vision with applications in caption evaluation, hallucination detection, data curation, and the benchmarking of text-to-image (T2I) generators.

Physical fidelity has received increasing attention in world models and video generation, yet how video representations encode physical information remains less understood.

Autoregressive (AR) video models excel at causal generation, but their reliance on next-chunk prediction confines them to a short-sighted, reactive paradigm.

Masked diffusion language models (dLMs) offer a promising parallel alternative to autoregressive models for complex reasoning.

LLM-guided evolutionary methods, such as AlphaEvolve, have emerged as powerful approaches for challenging computational optimization problems, such as circle packing.

Modern chess engines are silent experts: they play at a superhuman level, but do not offer explanations for their play.

Inspired by human vision, we introduce a framework using active gaze to enable fine-grained bimanual manipulation with only a single stereo camera.

We introduce 4DCodeBench, a benchmark for 4D inverse graphics through code generation, in which agents reconstruct dynamic scenes from video as executable graphics programs.

A benchmark score compares a system output against a reference, and methodological attention falls almost entirely on the first term.

We investigate a general-purpose layer for deep learning that, instead of compressing arbitrary-size training data into fixed-size weight matrices, stores a new pair of key-value representations for every data point duri…

AI agents are becoming increasingly capable of generating scientific code, but generating code is not the same as improving the algorithms behind it.

Autoregressive transformers remain comparatively weak for protein sequence and structure generation.

A trainable input embedding table assigns each vocabulary item an independently adjustable vector.

Language-model systems can separate contextual computation, persistent storage, and exact execution instead of updating all capabilities through one shared parameter system.

A language model reads long text in one quadratic forward pass, stops at the context window, and loses accuracy with length before reaching it.