AI research papers.

Reinforcement learning with verifiable rewards (RLVR) is reshaping how large language models (LLMs) acquire reasoning capabilities, yet the optimizers used remain largely inherited from teacher-forcing training.

Reconstructing reference videos as editable dynamic 3D scenes is an important path toward controllable video generation for filmmaking: these scenes provide modifiable, renderable visual proxies that enable explicit cont…

Processing multilingual text requires both identifying a broad range of languages and distinguishing the language of each token in inputs containing multiple languages.

Relational learning typically predicts from a database state, but decision making asks a different question: how will that state change under an action?

Generating and editing presentation slides requires models to combine accurate text rendering, instruction following, visual grounding, aesthetic design, and precise editing within one canvas.

Proximal Policy Optimization (PPO) is widely used in reinforcement learning with verifiable rewards for language-model reasoning.

Chest X-ray (CXR) interpretation often requires integrating visual findings with task-specific evidence rather than relying on direct visual recognition alone.

Vision-language-action (VLA) models show strong potential for end-to-end autonomous driving by leveraging world knowledge and understanding to excel in long-tail scenarios.

Evaluating model uncertainty on real-world data is challenging because each instance provides only a single observed label, leaving true conditional class probabilities unknown.

Knowledge-based Visual Question Answering (VQA) requires reasoning that extends beyond visible image content, demanding external commonsense and factual knowledge.

Existing interpretability tools do not test whether a representation keeps the same content when only the phrasing changes.

Evaluation has been central to advances in AI, and it is equally critical to autonomous discovery.

Linear-time sequence models typically store long-context information in a recurrent state.

Multi-player interactive world models aim to simulate shared environments from multiple players' perspectives, but achieving robust generation while scaling to more players remains challenging.

Generating images beyond the training resolution of Diffusion Transformers (DiTs) remains challenging, as the number of tokens grows rapidly with resolution, full attention becomes expensive, and outputs often suffer fro…

Can a vision-language model tell left from right?

Generative discovery in science requires coupling high-level scientific reasoning with the ability to generate samples respecting the structure of a design space.

We introduce StartLux-Decision, decision models from 0.8B to 27B parameters that combine efficient single-pass prediction with adaptive, confidence-guided reasoning.

Removing audio can shift a visual predictor's output without showing whether the task uses the intended audio information.

Generating complete 3D assets from sparse, unposed images requires preserving observed geometry while completing unseen regions.

Rapid advances in large language models have made software engineering agents capable of tackling complex coding tasks.

In embodied manipulation, Vision Language Action (VLA) models can execute atomic skills, but long-horizon manipulation requires recombining these skills without demonstrations for every skill combination.

Language models are becoming infrastructure, and infrastructure built in the open, by public institutions, on documented data, is the alternative to dependence on a few providers; Apertus shows this is viable at scale.

Linear attention promises the efficiency needed for long-context language modelling, yet its fixed-size recurrent state must simultaneously support ongoing computation and preserve information from an ever-growing histor…

Agent skills encode operational procedures and domain expertise as reusable knowledge assets, yet their distribution in plaintext exposes them to unauthorized copying and leakage.

Large language model agents are emerging as an important new class of users of desktop applications.

Recent focus in agentic reinforcement learning (RL) seeks to improve credit assignment in long-horizon tasks.

Vision-language navigation requires agents to follow natural-language instructions to reach target locations, making high-quality language supervision essential.

Large language model (LLM) agents are increasingly deployed as teams, yet they fail in characteristically social ways: information holders stay silent, dominant speakers anchor groups on wrong answers, and norm violation…

Reusable skills improve multimodal agents without updating model weights, but their effectiveness depends on both the evidence presented and the procedure used to read it.

Agent environments are essential for large language model (LLM) development.

We study pretraining when compute is abundant but unique training data are finite, requiring repeated training on the same data.

arXiv:2609.15066v2 Announce Type: replace-cross Abstract: We present Salesforce Koa, an enterprise language model built by post-training the open-weight Nemotron-3-Super-120B foundation model with reinforcement learning…

arXiv:2610.11167v1 Announce Type: cross Abstract: Vertical-domain few-shot classification remains challenging for small language models, as limited supervision makes it difficult to acquire domain-specific decision knowl…

arXiv:2610.12463v1 Announce Type: cross Abstract: In 2026, cybersecurity evaluations involving OpenAI, Anthropic, and Google agents reached real systems outside their authorized test scope.

Camera trajectories control viewpoint changes in video generation, scene reconstruction, and robotic perception.

Asynchronous rollout improves the efficiency of reinforcement learning with verifiable rewards (RLVR) by reducing under-utilization during long-tail generation.

Multi-robot systems (MRS) offer significant advantages over single-robot platforms by enabling information sharing and coordination for different tasks in complex environments.

Determining whether an image is AI-generated does not establish whether the situation it depicts is factually supported.

Large language models are increasingly deployed in applications that act on users' long-term behavior histories, from personal assistants to recommendation and advertising systems.

Latent diffusion models enable high-quality video generation, but video VAE encoding and decoding remain computationally expensive.

World Action Models (WAMs) commonly couple video and action experts through dense, layer-wise Mixture-of-Transformers (MoT) attention, exposing action prediction to irrelevant visual features while enforcing matched expe…

AI coding agents now write much of the code that ships, yet whether that code is secure remains unclear.

Large language models (LLMs) are increasingly expected to orchestrate diverse tools over long-horizon tasks.

Language models are increasingly used to create synthetic survey samples, also called silicon samples.

Audio source separation models must preserve fine spectral structure while keeping the cost of the separator tractable.

Recent bidirectional encoders have been modernized with innovations from causal large language models (LLMs).

Contrastive learning has become the standard paradigm for multimodal representation learning, aligning semantically matched pairs while separating mismatched ones in a shared embedding space.

High-fidelity closed-loop simulation for autonomous driving increasingly relies on dense multi-view 3D reconstruction.

Foundation models need to support diverse inference budgets, yet training a separate model for each budget duplicates pretraining effort.