AI research papers.

Speculative decoding accelerates large language model inference through parallel verification of tokens proposed by a drafter.

Energy-based models have shown promising properties for domain-specific tasks but remain challenging in modeling general-purpose scenarios.

Web agents complete user requests by reading and acting on pages that third parties write, so an instruction planted on a page can redirect the agent away from the user's goal.

Jennifer Neville did not want to go into computer science—but that’s exactly where she landed.

OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.

Learn how OpenAI and Ironclad are training and evaluating AI agents on complex contracting workflows to advance computer use for professional work.

A safe action is not necessarily a viable one.

Proactive AI assistants continuously observe a user's activity and decide whether to provide new guidance or remain silent.

Static-analysis checker synthesis requires agents to interpret a defect specification, inspect a repository, implement analyzer-specific logic, and refine the checker through repeated compilation and analysis feedback.

Modern recommendation is shifting from platform-centric personalization toward user-governed personalization, where a personal LLM agent can act on the user's behalf across services.

State tracking requires an update at every input, but the depth a Transformer applies to each token is fixed regardless of sequence length.

Automated research systems increasingly run LLM agents over long horizons, but more inference does not by itself produce more progress: agents replay growing histories, duplicate one another's work, or stop experimenting…

Tool-augmented speech assistants typically serialize automatic speech recognition, large language model inference, and external tool execution.

Speculative decoding accelerates autoregressive generation in large language models.

As human--AI interactions become more conversational, full-duplex speech language models capable of natural real-time dialogue are growing in importance.

Typed decision models answer a declared question without generating text: a decision head returns a probability for each of the declared options in a single forward pass.

Tool-using agents make consequential changes to external state, yet correct outcomes do not guarantee that their actions were supported by evidence established beforehand.

Reinforcement learning (RL) for post-training large language models (LLMs) incurs substantial computation and memory overhead during rollout generation, which motivates low-precision rollout for efficient RL training.

A central capability of embodied agents is to accomplish complex objectives through sequences of interdependent tasks.

Large language models (LLMs) equipped with terminal access have demonstrated strong capabilities in automating software engineering tasks.

A central challenge for scientific agents is to turn analytical experience into reusable expertise grounded in physical evidence.

On-policy distillation (OPD) has been widely studied as a post-training method in which a student model obtains token-level supervision from a stronger teacher on its own rollouts.

Repository-level software engineering (SWE) is a challenging long-horizon setting: agents must reason over extended interactions, use tools, and adapt to stateful environments.

As large language model (LLM) agents become increasingly autonomous, they may pursue task performance through deception, raising concerns about their reliable deployment.

Scaling robotic foundation models requires diverse training data and reliable evaluation environments.

Multimodal large language models have become the dominant paradigm for visual understanding, but incur substantial costs by encoding inputs into dense, fixed-size patch tokens.

LLM agents often lack the operational knowledge to act reliably in new environments, as they must discover specific tool behaviors or environment conventions on their own.

Reinforcement learning with verifiable rewards (RLVR) turns agent experience into learning signals primarily through scalar outcome rewards after interaction.

Sensors are useful not only for understanding the world but also for deciding what to do next.

Agentic reinforcement learning (RL) disaggregates training from rollout, so each policy update must reach the rollout clusters before the next batch.

Improving embodied agents often focuses on optimizing the underlying model through training, while the surrounding agent harness that controls planning, context, and tool use is typically engineered.

On-Policy Distillation (OPD) trains a student on its own generations using teacher feedback.

Long-context inference and Retrieval-Augmented Generation (RAG) handle evidence selection at vastly different scales, from a single long prompt to an entire corpus.

Recent game design agents have made substantial progress in generating playable games.

Recently Large Language Models (LLMs) and LLM-based agents increasingly need to incorporate knowledge acquired after pretraining, e.g., domain facts, user preferences, documents, and interaction experience.

Assistants from 2011 answered and waited, and agents from 2023 did a task and stopped.

Recent advances in world models have enabled increasingly realistic visual synthesis.

Self-improving policies continually expose new failure patterns, changing what their judges must be able to verify.

Interactive video world models need to generate each video chunk efficiently while responding faithfully to user controls.

Large language models (LLMs) have become increasingly capable problem solvers, but being able to solve a problem is not the same as being able to teach it.

Current video editors can insert objects but often struggle to make them participate in interactions such as being picked up or manipulated.

Single-image scene generation aims to produce a complete 3D scene mesh from a single image, including surfaces the camera did not observe.

Video world models can produce visually convincing yet physically inconsistent sequences, raising concerns about their reliability for prediction and planning in embodied AI systems.

A declared task can sometimes certify a reduced source before a downstream encoder, codebook, rate, distortion target, or optimizer is chosen.

AI agents increasingly operate in environments where they can diagnose failures and improve through experience, yet existing evaluations largely measure what an agent can do at a fixed point in time rather than how effec…

Asynchronous reinforcement learning (RL) for large language model (LLM) agents trains one policy on trajectories generated by another: rollouts come from stale checkpoints, and the inference engine's probabilities differ…

Astra can act, yet reliable manipulation depends on the system through which it observes and controls the world.

Vision-language-action (VLA) models map visual observations and language instructions to continuous robot actions.

Human videos provide rich motion targets for humanoid learning, yet visually plausible references can still produce persistent failures under physics-based execution.

The rapid growth of 3D Gaussian Splatting (3DGS) research demands significant effort to reimplement papers before building on them.