arXiv 论文速递

3D Gaussian avatars support fast rendering, however, their real-time animation is often challenged by the costly neural inference. We address this bottleneck and show that the anim…

填补LLM网络安全工具调用评测空白，提出细粒度基准与免运行时可验证奖励，推动安全智能体落地。

LLMs are increasingly applied to cybersecurity workflows, where they are expected to translate analysts' intent into tool invocations. However, existing evaluations focus on knowle…

提出引导式自改进框架，减少具身智能体技能开发与奖励设计的人力依赖，提升跨任务机器人能力。

Building reliable robot capabilities across diverse tasks requires substantial human effort to develop and maintain skills, design rewards, and integrate perception with control. W…

揭示扩散Transformer中条件嵌入在每步去噪被重复使用的问题，提出嵌入预测改进生成质量，为扩散模型条件机制提供新思路。

In diffusion transformers, a class label or a text prompt is embedded once, and the same condition is reused at every denoising step. We ask whether predicted embeddings can serve …

What makes great scientists great? Even as AI systems start to make progress on open problems, scientists remain far ahead of them at sensing which prior idea, buried in an ever-gr…

High-resolution 3D generation increasingly relies on voxel latents and multi-stage pipelines that first predict active structure and then synthesize local geometry. While effective…

We show that multimodal models possess strong reasoning abilities and that an appropriate harness can unlock their potential to solve tasks across diverse interactive environments.…

Full-parameter fine-tuning of large language models (LLMs) incurs substantial optimizer state memory overhead, limiting the model sizes that fit on modern GPUs. Existing approaches…

Several state-of-the-art methods for online reinforcement learning in continuous control improve policies using action gradients of a learned critic. However, critics are typically…

The generalized Schrödinger bridge on a graph moves mass between two distributions while charging a cost for the states visited. It has been approached by learning the rates of a c…

针对离散扩散语言模型在双向推理上的局限，提出分层连续扩散框架，为文本生成提供超越自回归的新范式。

Discrete diffusion language models offer a compelling alternative to autoregressive generation for tasks demanding bidirectional reasoning and global constraint satisfaction. Yet t…

诊断LLM数学推理缺失的基础能力并给出修复方案，对理解与提升大模型推理可靠性有重要价值。

While Large Language Models (LLMs) have demonstrated striking capabilities on frontier mathematical problems, it remains unclear whether they possess the structural mathematical un…