独立 AI 资讯库
从自动抓取数据库实时读取，适合快速浏览、检索和后续扩展成完整资讯站。
Dragonfly v2.5.0 现已发布，具体更新内容包括： 新功能和增强 从 Hugging Face 和 ModelScope 直接下载仓库 Dragonfly Client 现在支持直接从 Hugging Face 和 ModelScope 下载模型仓库。用户可以运行如下命令来获取仓库：dfget hf://deepseek-ai/DeepSeek-OCR和 dfget modelscope://models/deepseek-ai/DeepSeek-OCR。...
千问宣布正式发布实时语音交互对话模型 Qwen-Audio-3.0-Realtime——一个既要毫秒级响应、又不牺牲推理深度的实时语音模型。 公告称 ，该模型在智商、Agent工具调用、共情对话、双工交互流畅度四条主线上同步升级，力求"又快又聪明"。模型提供推理更强的 Plus 版本与速度更快的Flash版本，适用于智能客服、教育培训、娱乐...
RAG 让 LLM 从“闭卷考试”变为“开卷考试”——先检索相关资料，再基于资料生成答案。本文用 LangChain 从零搭建一个完整的 RAG 系统，涵盖 Document 设计、Embedding
RAG 最常见的用途，是让大模型回答它原本不知道，或者不能只靠自身记忆准确回答的问题。 企业内部文档、产品手册、课程资料、小说原文、项目代码说明，都可以成为 RAG 的外部知识来源。这些内容不一定存在
kkRepo 是什么？ kkRepo 是一款社区驱动、完全开源的自托管制品仓库，旨在解决 Sonatype Nexus 社区版的各类限制与痛点，为社区提供开放、可靠且可持续演进的制品管理方案。目前已支持 Maven 、npm 、PyPI 、Go 、Helm 、Cargo/Rust 、Dart/Pub 、Docker/OCI 、NuGet 、RubyGems 、Yum 和 Raw 等制品格式。 功能特性 支持...
如果你正在准备 AI 应用开发方向的面试，RAG（Retrieval Augmented Generation，检索增强生成）这个词你一定绕不开。但市面上的教程要么堆概念不讲代码，要么直接丢一坨
简介 SLF4J 是 Java 里很常见的日志门面。 门面这两个字听起来有点抽象，放到日志里理解就很简单：业务代码只调用 SLF4J 的 API，真正写日志的工作交给 Logback、Log4j2、J
Explore GPT-Red, OpenAI’s automated red teaming system that uses self-play to improve AI safety, alignment, and prompt injection robustness.
点击查看原文>
Claude Code 高效使用指南：让 AI 真正完成任务，而不只是生成代码 一句话结论 Claude Code 的 Prompt Library 不是“万能咒语合集”，而是一套围绕软件交付生命周期
告别 AI 幻觉！LangChain RAG 完全实战：Document + MemoryVectorStore 打造轻量级语义知识库 前置知识 了解 Embedding（向量化）的基本概念 了解 c
Google 推出新版图片搜索功能，Spotify 开启语音交互功能内测等。查看全文
The funding discussions point to investor interest in applying AI to make breakthroughs in life sciences.
June 2026 is about visibility and trust with a clearer view of your GitHub Copilot usage, a new trust layer for MCP servers, and the first C++ scenarios for the… The post GitHub Copilot in Visual Studio — June update appeared first on The GitHub Blog.
arXiv:2607.12304v1 Announce Type: new Abstract: A score on a temporal video question answering benchmark is meant to measure that a model has temporal understanding, but it conflates two questions. 1. The task question: is the question even temporal, does it need several frames and their order? and
arXiv:2607.12293v1 Announce Type: new Abstract: Predicting pedestrian crossing intention is a safety-critical task for autonomous driving, yet existing approaches often rely on single-modal inputs or dense multimodal fusion strategies that inadequately capture complementary visual and kinematic inf
arXiv:2607.12278v1 Announce Type: new Abstract: Recent vision-language models (VLMs) for computational pathology report striking zero-shot performance on whole-slide image (WSI) visual question answering (VQA) benchmarks. We audit these claims and find them fundamentally compromised by data leakage
本文提出了一种受治理的多智能体架构，使智能体能够递归自我改进，并引入“个人奇点”作为有边界的人机共同发展目标，帮助用户在有选择的领域内扩展能力。每个智能体由目标契约、范围、工具注册表及端到端基准等要素定义，超出范围的任务将移交给其他负责或新创建的细分智能体。架构融合了快速规划-执行-验证循环与较慢的证据把关改进循环，并配备了安全不变量、基准设计与实现路线图。本文为立场与系统设计论文，并非宣称无限制的递归自我改进或个人奇点已经实现。
arXiv:2607.12176v1 Announce Type: new Abstract: Ambivalence and hesitancy (A/H) undermine digital behaviour-change interventions, and recognizing them automatically from video is the goal of the ABAW A/H challenge on the BAH dataset. We describe our system for the 11th edition of the challenge: a c
arXiv:2607.12076v1 Announce Type: new Abstract: We present NEEDL-Bench, a microscopy detection benchmark for Swiss Needle Cast (SNC), a fungal disease of Douglas-fir trees. Douglas-fir is a keystone species of major ecological and economic importance as a softwood timber resource, and SNC affects p
一篇发表于 arXiv (2607.12042v1) 的新论文提出了 SymbOmni 框架。该研究指出当前整体式视觉生成模型缺乏累积学习与进化能力，并试图通过符号概念学习来演进智能体全能模型。这项工作可能为多模态交互与持续学习的 Agent 开发提供新思路。
arXiv:2607.11939v1 Announce Type: new Abstract: Accurate pedestrian trajectory prediction in crowded environments remains challenging due to the multimodal uncertainty of human motion and the variable complexity of motion dynamics across different scene contexts. Existing goal-conditioned models re
本文提出PFAdapter框架，通过层次化LoRA分解实现个性化联邦学习中的多模态大模型高效适配。针对分布式网络边缘环境下的协作学习与数据隐私需求，该方法在保持模型性能的同时显著降低通信开销。该工作为构建联邦化、个性化的智能代理系统提供了新的参数高效微调范式，对边缘AI、通信网络及隐私保护应用具有重要参考意义。
arXiv:2607.12067v1 Announce Type: new Abstract: Institutional equity holdings disclosed in SEC Form 13F filings provide a rich temporal record of portfolio decisions by large investment managers. However, forecasting future allocations and modeling future demand remains challenging due to disclosur
arXiv:2607.11976v1 Announce Type: new Abstract: Indexer-TopK, the operation to compute the scores and select the top-k candidates, is widely used by sparse attention kernels in large language models and vector retrieval in recommendation systems and vector databases. However, existing GPU-based Ind
arXiv:2607.11953v1 Announce Type: new Abstract: Does a reinforcement-learning agent that earns high reward represent its task's latent state, or only a reward-correlated shortcut? The question is usually unanswerable: the "true state" is undefined. We make it exactly answerable with a white-box ins
arXiv:2607.11928v1 Announce Type: new Abstract: Single-shot fringe projection profilometry (FPP) networks that regress depth directly can exploit a shape-prior shortcut, recovering depth from object boundaries rather than from fringe phase. On a photorealistic synthetic benchmark (15,600 fringe ima
arXiv:2607.11897v1 Announce Type: new Abstract: Linear attention replaces softmax attention's growing KV cache with a fixed recurrent state, but this compression limits exact state tracking and long-context memory. We introduce \emph{Semidirect Fourier Delta Attention} (SFDA), a phase-controlled ge
arXiv:2607.12395v1 Announce Type: new Abstract: Reinforcement learning with verifiable rewards without human-annotated data, often referred to as zero RL, has emerged as a powerful paradigm for eliciting chain-of-thought reasoning. However, due to computational constraints, existing studies are lar