T
traeai
Sign in

公司

什么是 KDnuggets

也叫:kd nuggets

提供数据科学和人工智能相关资讯的网站。

为什么现在值得关注?

最近变化

2026-06-09 · FastAPI适合构建高性能API,支持自动API文档生成。

KDnuggets 被反复提及时,通常意味着它正在影响产品路线、开发者工作流或 AI 产业判断。这个页面把分散材料合并成一个可持续更新的观察入口。

📰 KDnuggets 最新动态

已收录 11 篇与「KDnuggets」相关的 AI 资讯和分析。

KDnuggets 图标

Why Do LLMs Corrupt Your Documents When You Delegate?

KDnuggets1110 字 (约 5 分钟)
85

大型语言模型在多次交互中可能悄悄损坏用户委托编辑的文档,即使是最先进的模型如GPT-5也会出现内容损坏。

入选理由:最先进模型如GPT-5在20次交互后可能损坏25%的文档内容。

FeaturedArticle#LLM#文档编辑#AI#数据完整性英文
KDnuggets 图标

10 GitHub Repositories for Web Development in Python

KDnuggets1731 字 (约 7 分钟)
85

本文推荐了10个用于Python Web开发的GitHub仓库,涵盖API构建、全栈应用、仪表盘和机器学习演示等。

入选理由:FastAPI适合构建高性能API,支持自动API文档生成。

FeaturedArticle#Python#Web开发#GitHub#框架英文
A Gentle Primer on LLM Explainability

A Gentle Primer on LLM Explainability

KDnuggets772 字 (约 4 分钟)
85

LLM explainability is shifting from static to dynamic, multidimensional frameworks combining statistical methods and lightweight proxy models to enhance transparency and enable observable, debuggable AI systems in industry.

入选理由:SMILE框架通过局部输入扰动分析,使用统计距离测量生成热力图,揭示LLM输出的关键影响词。

FeaturedArticle#LLM#XAI#Explainability#SMILE#gSMILE英文
Time-Series Feature Engineering with Python Itertools

Time-Series Feature Engineering with Python Itertools

KDnuggets3263 字 (约 14 分钟)
85

Use Python's itertools module to build time series features with flexible iteration methods.

入选理由:文章介绍了如何利用 itertools 构建七类时间序列特征。

FeaturedArticle#Python#Time Series#itertools英文
5 Must-Know Python Concepts

5 Must-Know Python Concepts

KDnuggets1324 字 (约 6 分钟)
85

Mastering these five core Python concepts can significantly improve code efficiency and maintainability.

入选理由:列表推导式比循环快,生成器表达式节省内存。

FeaturedArticle#Python#Programming#Development英文
How to Build Vector Search From Scratch in Python

How to Build Vector Search From Scratch in Python

KDnuggets1886 字 (约 8 分钟)
85

This article explains how to build a vector search system from scratch using Python and NumPy, demonstrating the storage, normalization, and cosine similarity calculation of embedding vectors.

入选理由:使用NumPy构建向量搜索系统

FeaturedArticle#Python#Vector Search#Machine Learning中文
3 SpaCy Tricks for Efficient Text Processing & Entity Recognition

3 SpaCy Tricks for Efficient Text Processing & Entity Recognition

KDnuggets2276 字 (约 10 分钟)
83

By selectively loading pipeline components, parallel batching, and combining rule‑based with statistical NER, spaCy’s text processing speed can be increased 2–3× while reducing memory usage.

入选理由:排除不必要的组件(如 parser、tagger)可将 1,000 条文本的 NER 处理时间从 2.85 秒降至 1.12 秒,提升 2.5×。

FeaturedArticle#spaCy#NLP#TextProcessing#EntityRecognition#PerformanceOptimization中文
What the Agentic Era Means for Data Science

What the Agentic Era Means for Data Science

KDnuggets1505 字 (约 7 分钟)
82

Data science has entered the 'Agentic Era,' shifting the focus from manual procedural execution to the evaluation and supervision of autonomous AI agents. These agents use a 'Perceive-Reason-Act-Evaluate' loop to handle data cleaning, EDA, and tuning, evolving the data scientist's role from an executor of 'how' to a decision-maker of 'whether'.

入选理由:AI 智能体采用迭代循环机制(感知-推理-行动-评估),而非传统的单次 Prompt 响应模式。

FeaturedArticle#AI Agents#Data Science#LangGraph#AutoGen#LLM Orchestration英文
5 Must-Know Python Concepts for Data Scientists

5 Must-Know Python Concepts for Data Scientists

KDnuggets2705 字 (约 11 分钟)
82

This article introduces five essential Python concepts for data scientists, emphasizing NumPy vectorization and broadcasting mechanisms that significantly improve data processing performance, showing up to 26x speedup compared to traditional loops.

入选理由:使用NumPy向量化可将数组运算速度提升至传统Python循环的26倍以上

FeaturedArticle#Python#Data Science#NumPy#Vectorization#Performance英文
System Design Interview Questions: A Handy Collection

System Design Interview Questions: A Handy Collection

KDnuggets1383 字 (约 6 分钟)
75

System design interview skills remain irreplaceable in the AI era; this article collects 10 excellent GitHub open-source repositories to help engineers prepare for system design interviews, covering comprehensive learning paths from beginner guides to practical questions.

入选理由:系统设计技能因涉及权衡决策和工程判断而难以被AI替代

FeaturedArticle#System Design#Interview Preparation#GitHub#Engineer英文
10 GitHub Repositories to Master FastAPI

10 GitHub Repositories to Master FastAPI

KDnuggets1381 字 (约 6 分钟)
75

The article recommends 10 GitHub repositories to help developers master the FastAPI framework through real projects.

入选理由:提供10个真实项目学习FastAPI

FeaturedArticle#FastAPI#GitHub#Python#API Development中文

与「KDnuggets」经常一起出现的 AI 术语。

💡 想追踪「KDnuggets」的长期趋势?去 实体雷达 · KDnuggets 查看详细分析和跨材料问答。

AI may generate inaccurate information. Please verify important content.