系列第 ③ 篇，对应技能地图第 2 格「Grounding」。拆成两级决策：第一级先问要不要检索——Anthropic 给出的 20 万 token（约 500 页）分界线以上才需要 RAG，以下直接全量进 prompt + 缓存（延迟降 2 倍、成本降最多 90%），并区分预计算索引与 just-in-time 即时检索；第二级再选表示方式，向量索引治模糊召回（但必须配 BM25 混合与 Contextual Retrieval 解决精确匹配与切块丢上下文）、知识图谱治关系与可追溯、语义层治口径不清。附可量化收益表（检索失败率 5.7% → 3.7% → 2.9% → 1.9%）、四个实现注意项、context rot 与上下文压缩/笔记/子智能体三件套，以及一张可抄的选型决策树。

Late-interaction retrieval is the state-of-the-art for visual document search, but it pays for its accuracy in storage. …

Large language model (LLM) agents can benefit from reusable skills distilled from prior task experience, yet existing sk…

Text-to-motion (T2M) generation maps natural language to human joint movements, aiding gaming, VR, and robotics. Retriev…

Multimodal entity linking grounds entity mentions in text and images to knowledge-base entries. These systems degrade on…

Large language model (LLM) agents increasingly rely on external skills, but routing user requests over large skill regis…

Robot learning increasingly depends on broad and diverse demonstrations, yet collecting robot data remains expensive and…

Benchmark researchers and developers of large language models (LLMs) and other AI systems need to find relevant evaluati…

把 2026 年可核查的公开证据整理成一张六层能力路线图——预训练、后训练 RL、推理时计算、上下文与记忆、智能体与 Harness、世界模型。含 Meta ScaleRL 40 万 GPU 小时实验结论、RL 预算占比 10%–30% 口径、Chinchilla 对比、Meta-Harness 6x 差距等数据锚点，并给出优先级表与算法工程师/产品经理的行动建议。

Video understanding has rapidly evolved toward video large language models (VideoLLMs): systems that couple video repres…

Ensuring the safety of autonomous driving is a critical challenge. Scenario-based testing is a systematic process used t…

Streaming video understanding is a critical capability for real-world applications, including embodied intelligence, aut…

Long-horizon manipulation is partially observable: the information needed to choose the next action may appear only in o…

检索增强生成的核心流程：向量化、检索、重排与生成