Few-step streaming audio--video generation requires both causal modeling and step distillation, yet standard training recipes face two context-related challenges. Teacher forcing p...
An agent that uses tools typically responds to what the user explicitly asks, yet completing the task may require information the user never requested. Work on proactive agents mai...
High-resolution video generation is expensive, as its cost grows rapidly with the number of spatiotemporal tokens. A practical alternative first generates a lower-resolution video ...
Policies with similar mean returns can differ sharply in rare failures, yet estimating lower-tail conditional value-at-risk (CVaR) accurately can require many costly rollouts. When...
Mixture-of-Experts (MoE), popularized by large language models, is a promising paradigm for scaling visual generative models. However, conventional token-wise MoE routes tokens ind...
Optical Character Recognition (OCR) is evolving from plain-text transcription toward general visual intelligence, requiring models to recognize, localize, and reason over textual i...
Emotion-conditioned text-to-speech (TTS) models may fail to express the requested emotion reliably, and improving controllability by additional training is costly in both computati...
Answering questions and completing tasks over large document collections often requires connecting evidence spread across multiple documents, such as a project's approval recorded ...
Existing online video segmentation methods struggle to track objects in long, complex videos with long-term occlusions. We hypothesize that this limitation is caused by (i) the ina...
Recent efforts to scale tool-use post-training have largely centered on the synthesis of executable environments, which constitute only one component of a broader agentic interacti...
Historical street-view imagery records urban evolution, but uneven coverage leaves substantial gaps in historical records. Generating plausible past appearances requires restoring ...
System One models such as Jev offer an efficient alternative to generative language models for tasks that require decisions rather than open-ended responses. However, existing Jev ...
Skills are external artifacts that Large Language Models (LLMs) consume at inference time to improve their performance on specialized domains by incorporating relevant procedural a...
Recurrent reasoning models have attracted growing attention for scaling test-time computation, typically by iteratively refining latent states with shared parameters. However, thes...
A small trainable advisor can steer a frozen language-model executor using natural-language advice. In addition to learning from task rewards, the advisor can use feedback from com...
Existing Large Language Model (LLM) routing methods score LLMs independently to select top-k models. However, this ignores model correlations and enforces a rigid computational bud...
Agents can learn from past executions, but enabling different agents to reuse and build on one another's experience remains challenging. We introduce EpiCon, a shared multimodal me...
Estimating mutual information (MI) from samples is a central objective in a variety of scientific fields. Modern neural estimators are accurate in the large-data regime, but they f...
Manipulating next-token probabilities during generation can bypass the safety alignment of large language models. Existing approaches, however, rely on access to model weights or n...
Autonomous agents have made rapid progress in general-purpose computer use, but reliable automation of professional industrial engineering remains out of reach, as engineering work...
Large language model (LLM) routing aims to assign each query to the most suitable model from a heterogeneous candidate pool, improving the quality--efficiency trade-off of LLM infe...
Answering questions about long videos often requires connecting events involving the same objects across hours or days. Chronological descriptions and text-derived entities can lea...
Agent performance depends on both reasoning ability and the environment in which it acts. We study test-time AI-for-AI, asking how a Builder can learn to construct better execution...
Latent world models learn action-conditioned dynamics in representation space and often score candidate actions by Euclidean distance to a goal representation. Joint training typic...