用全局地图搭配局部视图，为多视角3D推理提供稠密奖励信号，突破空间理解瓶颈。
arXiv:2606.23557v1 Announce Type: new Abstract: Multi-view 3D Visual Question Answering (MV3D-VQA) requires integrating partial observations into a co…
告别对话失忆，用属性密集奖励让大模型真正记住历史，长期记忆构建有新招了。
arXiv:2601.05488v4 Announce Type: replace Abstract: Maintaining consistency in long-term dialogues remains a fundamental challenge for LLMs, as standa…
用密集奖励实现领域自适应强化学习，突破代码生成边界
arXiv:2605.21180v1 Announce Type: new Abstract: Large language models show strong potential for automated code generation, but lack guarantees for cor…