无需训练的多智能体系统新范式，结构化记忆引导自适应。
arXiv:2606.08702v1 Announce Type: new Abstract: Recent advances have improved the adaptive capabilities of LLM-based multi-agent systems (MAS) through…
无需额外训练，用现成大模型就能给数学推理过程打分，性能媲美专用过程奖励模型。
arXiv:2606.01682v1 Announce Type: cross Abstract: Selecting the best response from multiple small-model samples using a stronger scorer is a simple in…
无需训练的轻量级视觉推理方法LookWise，让多模态大模型更高效地“细看”图像细节
arXiv:2603.00171v3 Announce Type: replace-cross Abstract: Multimodal Large Language Models (MLLMs) are shifting towards "Thinking with Images" by acti…
无需训练的零样本组合图像检索新方法，通过语义转换与传输协作实现精准高效匹配。
arXiv:2605.21261v1 Announce Type: new Abstract: Training-free zero-shot composed image retrieval models are recently gaining increasing research inter…
无需训练即可校正扩散模型后验，解决高密度区域外采样难题的新方法。
arXiv:2507.05482v3 Announce Type: replace Abstract: Training-free diffusion guidance offers a flexible framework for leveraging off-the-shelf classifi…
无需训练的自适应注意力抑制方法，精准移除图像中指定对象并修复背景。
arXiv:2605.15921v1 Announce Type: new Abstract: Object removal aims to eliminate specified objects from images while plausibly inpainting the affected…