我的第二幻想乡,记录代码、计算机和生活的思考。https://lanxinmob.dev/https://lanxinmob.dev/posts/cs2857/https://lanxinmob.dev/posts/cs2857/回顾 CS285 的整体知识框架，并讨论深度强化学习仍然面临的挑战与开放问题。Tue, 20 Oct 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2856/https://lanxinmob.dev/posts/cs2856/从逆强化学习与特征匹配出发，延伸到 RLHF、序列模型，以及迁移学习与 Meta-RL。Wed, 16 Sep 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2855/https://lanxinmob.dev/posts/cs2855/从固定数据集中学习更好的策略：离线 RL、CQL、COMBO、Trajectory Transformer，以及离线策略评估与 RL 理论。Thu, 20 Aug 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2854/https://lanxinmob.dev/posts/cs2854/讨论稀疏奖励下的探索问题，包括深度强化学习中的乐观探索与后验采样方法。Tue, 11 Aug 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2853/https://lanxinmob.dev/posts/cs2853/梳理基于模型强化学习的核心方法，包括动力学模型学习、LQR、模型集成、随机射击、CEM、MBPO与GPS等Wed, 29 Jul 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2852/https://lanxinmob.dev/posts/cs2852/从策略梯度、自然梯度与重要性采样，到 Actor-Critic、GAE、DQN 和 SAC，梳理 model-free 强化学习的主线。Thu, 23 Jul 2026 18:30:00 GMThttps://lanxinmob.dev/posts/cs2851/https://lanxinmob.dev/posts/cs2851/主要探讨模仿学习的理论瓶颈、多模态连续动作表示，以及强化学习算法的基本结构与数学直觉。Mon, 20 Jul 2026 18:30:00 GMThttps://lanxinmob.dev/posts/ppo/https://lanxinmob.dev/posts/ppo/PPO,GRPOSat, 20 Jun 2026 09:00:00 GMThttps://lanxinmob.dev/posts/minimal/https://lanxinmob.dev/posts/minimal/应用视角的osThu, 18 Jun 2026 23:00:00 GMThttps://lanxinmob.dev/posts/whywhat-and-how-aobut-os/https://lanxinmob.dev/posts/whywhat-and-how-aobut-os/nju os 2026概述Mon, 01 Jun 2026 23:00:00 GMThttps://lanxinmob.dev/posts/simulation/https://lanxinmob.dev/posts/simulation/libero,robocasa and mujocoSat, 30 May 2026 18:30:00 GMThttps://lanxinmob.dev/posts/ysyx-day10/https://lanxinmob.dev/posts/ysyx-day10/f6完结Mon, 23 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day9/https://lanxinmob.dev/posts/ysyx-day9/f6Thu, 19 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day8/https://lanxinmob.dev/posts/ysyx-day8/f5sCPUWed, 18 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day7/https://lanxinmob.dev/posts/ysyx-day7/f5li指令Tue, 17 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day6/https://lanxinmob.dev/posts/ysyx-day6/f4Mon, 16 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day5/https://lanxinmob.dev/posts/ysyx-day5/f3计数器、电子时钟Sun, 15 Mar 2026 12:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day4/https://lanxinmob.dev/posts/ysyx-day4/f3锁存器、寄存器Sat, 14 Mar 2026 23:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day3/https://lanxinmob.dev/posts/ysyx-day3/f3原码、反码加法器Fri, 13 Mar 2026 23:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day2/https://lanxinmob.dev/posts/ysyx-day2/f3译码器、编码器Thu, 12 Mar 2026 23:00:00 GMThttps://lanxinmob.dev/posts/ysyx-day1/https://lanxinmob.dev/posts/ysyx-day1/f2和f3搭建门电路Mon, 09 Mar 2026 23:00:00 GMThttps://lanxinmob.dev/posts/the-future-of-ai-and-course-wrap-up/https://lanxinmob.dev/posts/the-future-of-ai-and-course-wrap-up/The Future of AI and Course Wrap-upFri, 06 Mar 2026 14:30:00 GMThttps://lanxinmob.dev/posts/decision-trees/https://lanxinmob.dev/posts/decision-trees/Decision TreesThu, 05 Mar 2026 14:30:00 GMThttps://lanxinmob.dev/posts/ml-intro/https://lanxinmob.dev/posts/ml-intro/Naive Bayes and PerceptronsWed, 04 Mar 2026 14:30:00 GMThttps://lanxinmob.dev/posts/vpi-and-hmm/https://lanxinmob.dev/posts/vpi-and-hmm/VPI and HMMSun, 01 Mar 2026 19:30:00 GMThttps://lanxinmob.dev/posts/bayes-nets-inference/https://lanxinmob.dev/posts/bayes-nets-inference/贝叶斯网络推理算法Sun, 01 Mar 2026 14:30:00 GMThttps://lanxinmob.dev/posts/%E6%A6%82%E7%8E%87%E8%AE%BA%E5%9F%BA%E7%A1%80%E4%B8%8E%E8%B4%9D%E5%8F%B6%E6%96%AF%E7%BD%91%E7%BB%9C/https://lanxinmob.dev/posts/%E6%A6%82%E7%8E%87%E8%AE%BA%E5%9F%BA%E7%A1%80%E4%B8%8E%E8%B4%9D%E5%8F%B6%E6%96%AF%E7%BD%91%E7%BB%9C/概率论基础与贝叶斯网络Mon, 23 Feb 2026 14:30:00 GMThttps://lanxinmob.dev/posts/%E8%B4%9D%E5%8F%B6%E6%96%AF%E7%BD%91%E7%BB%9C/https://lanxinmob.dev/posts/%E8%B4%9D%E5%8F%B6%E6%96%AF%E7%BD%91%E7%BB%9C/贝叶斯网络Mon, 23 Feb 2026 14:30:00 GMThttps://lanxinmob.dev/posts/mdp/https://lanxinmob.dev/posts/mdp/CS188 课程笔记：马尔可夫决策过程、策略与价值迭代，以及不确定性下的决策建模。Thu, 12 Feb 2026 14:32:00 GMThttps://lanxinmob.dev/posts/rl/https://lanxinmob.dev/posts/rl/CS188 课程笔记：无模型强化学习，涵盖 Q-Learning、近似 Q-Learning 与策略搜索。Thu, 12 Feb 2026 14:32:00 GMThttps://lanxinmob.dev/posts/cs1884/https://lanxinmob.dev/posts/cs1884/CS188 课程笔记：对抗搜索与博弈，涵盖 Game Tree、Minimax、Alpha-Beta 剪枝和随机性游戏。Thu, 12 Feb 2026 09:32:00 GMThttps://lanxinmob.dev/posts/iptable-note/https://lanxinmob.dev/posts/iptable-note/深入学习防火墙和iptables配置背后的五链四表Tue, 10 Feb 2026 09:00:00 GMThttps://lanxinmob.dev/posts/cs188/https://lanxinmob.dev/posts/cs188/CS188 Project 1 的搜索算法实现记录：DFS、BFS、UCS 与 A*。Sat, 07 Feb 2026 09:00:00 GMThttps://lanxinmob.dev/posts/react-note/https://lanxinmob.dev/posts/react-note/React 学习笔记：列表渲染与 key、数组 state 的更新，以及用 React 构建 UI 的思路。Sat, 17 Jan 2026 09:00:00 GMThttps://lanxinmob.dev/posts/unix-note/https://lanxinmob.dev/posts/unix-note/Unix 编程笔记：进程环境、fork/exec、进程终止以及僵尸和孤儿进程等第二章核心内容。Mon, 12 Jan 2026 09:00:00 GMThttps://lanxinmob.dev/posts/%E7%A8%8B%E5%BA%8F%E7%BA%BF%E7%A8%8B%E5%92%8C%E8%BF%9B%E7%A8%8B/https://lanxinmob.dev/posts/%E7%A8%8B%E5%BA%8F%E7%BA%BF%E7%A8%8B%E5%92%8C%E8%BF%9B%E7%A8%8B/从程序映像的内存分区出发，整理 Unix 进程、线程以及错误处理和调试的基础知识。Fri, 09 Jan 2026 09:32:00 GMThttps://lanxinmob.dev/posts/%E5%BC%82%E6%AD%A5%E4%B8%8E%E5%B9%B6%E5%8F%91/https://lanxinmob.dev/posts/%E5%BC%82%E6%AD%A5%E4%B8%8E%E5%B9%B6%E5%8F%91/从异步的核心特征出发，比较线程、进程、GIL、写时复制和进程间通信在并发中的取舍。Thu, 08 Jan 2026 23:40:00 GMThttps://lanxinmob.dev/posts/unix%E5%9F%BA%E7%A1%80/https://lanxinmob.dev/posts/unix%E5%9F%BA%E7%A1%80/Unix 基础笔记：头文件与编译链接、静态库和共享库、宏与条件编译，以及常用开发工具。Thu, 08 Jan 2026 09:32:00 GMThttps://lanxinmob.dev/posts/csps-advanced/https://lanxinmob.dev/posts/csps-advanced/CS188 课程笔记：CSP 进阶，包括高阶一致性、利用图结构优化和局部搜索。Wed, 31 Dec 2025 23:32:00 GMThttps://lanxinmob.dev/posts/cspi/https://lanxinmob.dev/posts/cspi/CS188 课程笔记：约束满足问题（CSP）的建模、约束图、标准搜索框架与求解思路。Wed, 31 Dec 2025 09:32:00 GMThttps://lanxinmob.dev/posts/stm32%E4%B8%B2%E5%8F%A3%E9%80%9A%E4%BF%A1/https://lanxinmob.dev/posts/stm32%E4%B8%B2%E5%8F%A3%E9%80%9A%E4%BF%A1/STM32 串口通信：CubeMX 配置、波特率与接线，以及轮询、中断、DMA 和空闲中断接收。Sat, 27 Dec 2025 01:32:00 GMThttps://lanxinmob.dev/posts/cs1882/https://lanxinmob.dev/posts/cs1882/CS188 课程笔记：边缘集、启发式搜索、Greedy 与 A*，以及图搜索的终止条件。Fri, 19 Dec 2025 09:32:00 GMThttps://lanxinmob.dev/posts/i2c%E9%80%9A%E4%BF%A1/https://lanxinmob.dev/posts/i2c%E9%80%9A%E4%BF%A1/I2C 通信协议基础：SDA/SCL 引脚、主从与半双工特性，以及起始、停止和应答等控制信号。Fri, 19 Dec 2025 01:32:00 GMThttps://lanxinmob.dev/posts/cs1881/https://lanxinmob.dev/posts/cs1881/CS188 课程笔记：从 AI 的定义和智能体分类出发，介绍搜索问题、通用树搜索与期望效用。Thu, 18 Dec 2025 23:00:00 GMThttps://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95week3/https://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95week3/CS61B 学习记录 week3：Java 的接口继承与实现继承，包括 overriding 和继承设计。Wed, 17 Dec 2025 21:34:00 GMThttps://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95week2/https://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95week2/CS61B 学习记录 week2：访问控制、嵌套类、泛型 AList 与 Java 包机制。Mon, 15 Dec 2025 23:00:00 GMThttps://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95/https://lanxinmob.dev/posts/cs61b%E5%AD%A6%E4%B9%A0%E8%AE%B0%E5%BD%95/CS61B 学习记录 week1：Java 的编译与解释流程、静态类型、static 语义和变量声明。Sun, 19 Oct 2025 23:00:00 GMThttps://lanxinmob.dev/posts/%E6%9B%B4%E6%8D%A2%E4%B8%BAnapcat%E6%A1%86%E6%9E%B6/https://lanxinmob.dev/posts/%E6%9B%B4%E6%8D%A2%E4%B8%BAnapcat%E6%A1%86%E6%9E%B6/记录从 Lagrange.OneBot 迁移到 NapCat 框架的安装、接入和运行过程。Wed, 15 Oct 2025 09:00:00 GMThttps://lanxinmob.dev/posts/%E6%9E%84%E5%BB%BA%E5%B9%B6%E5%8F%91%E5%B8%83%E5%88%B0pypi/https://lanxinmob.dev/posts/%E6%9E%84%E5%BB%BA%E5%B9%B6%E5%8F%91%E5%B8%83%E5%88%B0pypi/以及代码规范化操作Sun, 12 Oct 2025 23:00:00 GMThttps://lanxinmob.dev/posts/%E4%BA%BA%E7%A5%96/https://lanxinmob.dev/posts/%E4%BA%BA%E7%A5%96/《蛊真人》中的摘句：关于道路、选择与自我。Thu, 09 Oct 2025 09:00:00 GMThttps://lanxinmob.dev/posts/redis-note/https://lanxinmob.dev/posts/redis-note/常用redis指令Thu, 02 Oct 2025 09:00:00 GMThttps://lanxinmob.dev/posts/%E5%AE%89%E8%A3%85shellclash/https://lanxinmob.dev/posts/%E5%AE%89%E8%A3%85shellclash/在 Linux 或路由器上安装 ShellClash 的完整命令，包含 bash/sh 的选择、安装源、终端交互过程和安装后的管理方式。Mon, 29 Sep 2025 09:00:00 GMThttps://lanxinmob.dev/posts/mako_evolution/https://lanxinmob.dev/posts/mako_evolution/Mako-Bot（茉子）的开发记录：从聊天、提醒等基础功能，到记忆分层、图片与检索工具、Web 面板和自主行动。Thu, 25 Sep 2025 09:00:00 GMThttps://lanxinmob.dev/posts/java%E7%9A%84%E5%AD%A6%E4%B9%A0/https://lanxinmob.dev/posts/java%E7%9A%84%E5%AD%A6%E4%B9%A0/初识JavaWed, 17 Sep 2025 09:00:00 GMThttps://lanxinmob.dev/posts/git%E8%BE%A8%E6%9E%90/https://lanxinmob.dev/posts/git%E8%BE%A8%E6%9E%90/一些markdown语法和git指令辨析Sat, 13 Sep 2025 09:00:00 GMThttps://lanxinmob.dev/posts/sql/https://lanxinmob.dev/posts/sql/SQL 基础笔记：建表与约束、增删改查，以及在 Python 中操作数据库的常用写法。Sun, 10 Aug 2025 09:00:00 GMThttps://lanxinmob.dev/posts/method/https://lanxinmob.dev/posts/method/kalman HMM MCMTue, 05 Aug 2025 09:00:00 GMThttps://lanxinmob.dev/posts/server/https://lanxinmob.dev/posts/server/降临仪式圆满完成！Sun, 03 Aug 2025 09:00:00 GMThttps://lanxinmob.dev/posts/veseel/https://lanxinmob.dev/posts/veseel/翻阅的一些论文Sat, 02 Aug 2025 09:00:00 GMThttps://lanxinmob.dev/posts/nonebot/https://lanxinmob.dev/posts/nonebot/遇到的问题和收获Thu, 31 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/codex/https://lanxinmob.dev/posts/codex/李沐论文精读 DeepMind AlphaCode OpenAI CodexTue, 29 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/gpt/https://lanxinmob.dev/posts/gpt/李沐论文精读 GPT,-2,-3Mon, 28 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/bert/https://lanxinmob.dev/posts/bert/李沐论文精读 Pre-training of Deep Bidirectional Transformers for Language UnderstandingSun, 27 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/attetion-is-all-you-need/https://lanxinmob.dev/posts/attetion-is-all-you-need/李沐论文精读 TransformerFri, 25 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/data-label/https://lanxinmob.dev/posts/data-label/斯坦福实用机器学习1.4Fri, 25 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/resnet/https://lanxinmob.dev/posts/resnet/李沐论文精读 ResNetThu, 24 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/rethinking-generalization/https://lanxinmob.dev/posts/rethinking-generalization/interesting and not well understoodThu, 24 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/batch-file/https://lanxinmob.dev/posts/batch-file/paper link formatterWed, 23 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/alexnet/https://lanxinmob.dev/posts/alexnet/李沐论文精读 AlexNetTue, 22 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/capture-data/https://lanxinmob.dev/posts/capture-data/斯坦福实用机器学习1.2Tue, 22 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/automl/https://lanxinmob.dev/posts/automl/automlMon, 21 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/%E8%A7%84%E5%88%92/https://lanxinmob.dev/posts/%E8%A7%84%E5%88%92/总结和规划Sat, 19 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/less/https://lanxinmob.dev/posts/less/linuxFri, 18 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/oop/https://lanxinmob.dev/posts/oop/oopTue, 15 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/mysql/https://lanxinmob.dev/posts/mysql/MySQL索引原理Mon, 14 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/compete/https://lanxinmob.dev/posts/compete/lstm+mlp tpuSun, 13 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/request/https://lanxinmob.dev/posts/request/从 1ns 到 1s 建立对延迟数量级的直觉Fri, 11 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/n_t_d/https://lanxinmob.dev/posts/n_t_d/numpy tensor pandas Thu, 10 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/lstm/https://lanxinmob.dev/posts/lstm/练习rnn,lstm,完善网站Wed, 09 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/vm/https://lanxinmob.dev/posts/vm/虚拟化技术 训练测试集分割Tue, 08 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/npm/https://lanxinmob.dev/posts/npm/浴火重生Mon, 07 Jul 2025 09:00:00 GMThttps://lanxinmob.dev/posts/docker/https://lanxinmob.dev/posts/docker/跟着docker desktop的引导一步步学了点Sun, 06 Jul 2025 09:00:00 GMT