🌐 [translation-sync] Revise JAX intro lecture and add autodiff lecture by mmcky · Pull Request #16 · QuantEcon/lecture-python-programming.zh-cn

mmcky · 2026-04-08T00:26:54Z

Automated Translation Sync

This PR contains automated translations from QuantEcon/lecture-python-programming.

Source PR

#513 - Revise JAX intro lecture and add autodiff lecture

Files Added

✅ lectures/autodiff.md
✅ .translate/state/autodiff.md.yml

Files Updated

✏️ lectures/jax_intro.md
✏️ .translate/state/jax_intro.md.yml
✏️ lectures/numpy_vs_numba_vs_jax.md
✏️ .translate/state/numpy_vs_numba_vs_jax.md.yml
✏️ lectures/_toc.yml

Details

Source Language: en
Target Language: zh-cn
Model: claude-sonnet-4-6

This PR was created automatically by the translation action.

github-actions · 2026-04-08T00:27:34Z

✅ Translation Quality Review

Verdict: PASS | Model: claude-sonnet-4-6 | Date: 2026-04-08

📝 Translation Quality

Criterion	Score
Accuracy	9/10
Fluency	9/10
Terminology	9/10
Formatting	9/10
Overall	9/10

Summary: The translation is of high quality across all changed sections. It accurately conveys the technical content of the autodiff, JAX introduction, and NumPy/Numba/JAX comparison lectures. Terminology is consistent and follows established conventions. There is one minor spacing error in the vmap section, and a minor inconsistency between '操作' and '运算' in the Overall recommendations section versus the section headings. The mathematical content, code blocks, and MyST formatting are all correctly preserved. Technical terminology is consistently and correctly translated throughout all changed sections, including autodiff concepts, gradient descent, JIT compilation, and vmap. Mathematical expressions and LaTeX are fully preserved without any modifications. The translation of the 'How JIT compilation works' section accurately captures the technical pipeline description (trace → computational graph → XLA → optimized kernel). Code block structure, MyST directives, exercise/solution markup, and all formatting elements are correctly preserved across all added and modified sections. The 'Overall recommendations' section is translated fluently with appropriate academic register, accurately conveying the nuanced trade-offs between JAX, NumPy, and Numba.

Suggestions:

Vectorization with vmap section, paragraph 1: '它能自动将为单个输入编写的函数向量化' contains an extra space before '向量化'. Should be '它能自动将为单个输入编写的函数向量化'.
What is automatic differentiation? > ### Autodiff section: '导数通过链式法则将计算分解为各组成部分来构建' is slightly ambiguous. The English 'Derivatives are constructed by breaking calculations into component parts via the chain rule' could be more clearly rendered as '通过链式法则将计算分解为各个组成部分来构建导数' to preserve the subject-verb order more naturally.
Gradient Descent > ### A function for gradient descent: The docstring comments inside the code cell are left in English ('Function to be minimized', 'Extra arguments to the function', etc.). While code comments are often left untranslated, the surrounding prose is fully translated. This is acceptable but worth noting for consistency with the project's translation approach.
Overall recommendations section: '对于向量化操作' and '对于顺序操作' use '操作' while earlier sections consistently use '运算' (e.g., '向量化运算', '顺序运算' in headings). Should use '运算' for consistency: '对于向量化运算' and '对于顺序运算'.

🔍 Diff Quality

Check	Status
Scope Correct	✅
Position Correct	✅
Structure Preserved	✅
Heading-map Correct	✅
Overall	10/10

Summary: All three files are correctly modified with proper scope, position, structure, and heading-map updates reflecting the source changes.

This review was generated automatically by action-translation review mode.

Copilot

Pull request overview

This translation-sync PR updates the zh-cn JAX-related lectures to match upstream changes: it expands the JAX introduction (JIT internals, vmap, autodiff preview), adds a new autodiff lecture, and wires the new lecture into the book ToC.

Changes:

Added new lecture lectures/autodiff.md and its translation-sync state file.
Updated lectures/jax_intro.md with expanded sections (random keys, JIT pipeline explanation, vmap, autodiff preview) and new diagrams.
Updated lectures/numpy_vs_numba_vs_jax.md with minor code fixes/edits and a new “Overall recommendations” section; updated ToC/state files accordingly.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
lectures/autodiff.md	New autodiff lecture content (examples, gradient descent) added to the curriculum
lectures/jax_intro.md	Expanded JAX intro content (random keys, JIT internals, `vmap`, autodiff preview) + diagrams
lectures/numpy_vs_numba_vs_jax.md	Small code/wording fixes and new overall recommendations section referencing autodiff
lectures/_toc.yml	Adds `autodiff` to the book table of contents
.translate/state/autodiff.md.yml	Records translation sync state for the new lecture
.translate/state/jax_intro.md.yml	Updates translation sync state metadata
.translate/state/numpy_vs_numba_vs_jax.md.yml	Updates translation sync state metadata

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-04-08T00:31:41Z

lectures/autodiff.md

+kernelspec:
+  display_name: Python 3 (ipykernel)
+  language: python
+  name: python3


lectures/ 下的讲座文件通常在 frontmatter 中包含 translation: { title, headings } 映射（用于标题/目录同步与后续翻译更新）。autodiff.md 目前缺少该段，可能会导致目录标题或后续 translation-sync 工具无法正确对齐章节。建议补充 translation.title 以及主要小节的 translation.headings 映射。

Suggested change

name: python3

name: python3

translation:

title: Adventures in Autodiff

headings:

- - Overview

- 概述

Copilot · 2026-04-08T00:31:41Z

lectures/autodiff.md

+    while ϵ > tol and i < max_iter:
+        new_x = x - λ * df
+        new_df = f_grad(new_x, args)
+        Δx = new_x - x
+        Δdf = new_df - df
+        λ = jnp.abs(Δx @ Δdf) / (Δdf @ Δdf)
+        ϵ = jnp.max(jnp.abs(Δx))


grad_descent 的 Barzilai–Borwein 步长更新中分母 (Δdf @ Δdf) 可能为 0（例如梯度在相邻迭代间不变或接近收敛时数值下溢），从而产生 inf/NaN 并使后续迭代失效。建议在更新步长前对分母做零值/极小值保护（如保持旧 λ、回退到默认 λ 或提前停止）。

Copilot · 2026-04-08T00:31:41Z

lectures/autodiff.md

+x = jax.random.uniform(key, (n,))
+
+α, β, σ = 0.5, 1.0, 0.1  # Set the true intercept and slope.
+key, subkey = jax.random.split(key)
+ϵ = jax.random.normal(subkey, (n,))
+
+y = α * x + β + σ * ϵ


这里的注释把 α/β 的含义写反了：在 y = α * x + β + ... 中，α 是斜率、β 是截距。建议更正注释以避免读者混淆。

Copilot · 2026-04-08T00:31:41Z

lectures/jax_intro.md

+## 使用 `vmap` 进行向量化

-## 梯度
+JAX 的另一个强大变换是 `jax.vmap`，它能自动将为单个输入编写的函数 向量化，使其可以在批量数据上运行。


这句里 “函数向量化” 中间多了一个空格，影响排版与可读性。建议去掉多余空格。

Suggested change

JAX 的另一个强大变换是 `jax.vmap`，它能自动将为单个输入编写的函数向量化，使其可以在批量数据上运行。

JAX 的另一个强大变换是 `jax.vmap`，它能自动将为单个输入编写的函数向量化，使其可以在批量数据上运行。

Copilot · 2026-04-08T00:31:42Z

lectures/jax_intro.md

+# Root key
+ax.text(3, 3, "key₀", ha='center', va='center', fontsize=11,
+        bbox=box_style)
+
+# Level 1


图中标签使用了 Unicode 下标字符（如 key₀、key₁ 等）。在当前 i18n 字体设置下这些字符可能缺字导致渲染为方块，并且对屏幕阅读器也不友好。建议改为纯 ASCII（如 key0 / key_0）或使用 Matplotlib mathtext（如 $key_0$ ）以保证跨平台渲染一致。

Copilot · 2026-04-08T00:31:42Z

lectures/jax_intro.md

+
+ax.text(3, -0.5, "⋮", ha='center', va='center', fontsize=14)
+
+ax.set_title("PRNG Key Splitting Tree", fontsize=13, pad=10)


该图标题仍为英文（"PRNG Key Splitting Tree"），与整篇中文讲义的语言不一致，可能影响读者理解与检索。建议翻译为中文或在标题中同时提供中英文。

Suggested change

ax.set_title("PRNG Key Splitting Tree", fontsize=13, pad=10)

ax.set_title("PRNG 密钥拆分树（PRNG Key Splitting Tree）", fontsize=13, pad=10)

mmcky added 7 commits April 8, 2026 01:26

Update translation: lectures/autodiff.md

b90fa95

Update translation: .translate/state/autodiff.md.yml

08dc537

Update translation: lectures/jax_intro.md

0d81650

Update translation: .translate/state/jax_intro.md.yml

6a23bd3

Update translation: lectures/numpy_vs_numba_vs_jax.md

e01199d

Update translation: .translate/state/numpy_vs_numba_vs_jax.md.yml

ad7d562

Update translation: lectures/_toc.yml

36caf99

mmcky added action-translation automated labels Apr 8, 2026

Copilot AI review requested due to automatic review settings April 8, 2026 00:26

mmcky mentioned this pull request Apr 8, 2026

Revise JAX intro lecture and add autodiff lecture QuantEcon/lecture-python-programming#513

Merged

5 tasks

Copilot started reviewing on behalf of mmcky April 8, 2026 00:27 View session

Copilot AI reviewed Apr 8, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

🌐 [translation-sync] Revise JAX intro lecture and add autodiff lecture#16

🌐 [translation-sync] Revise JAX intro lecture and add autodiff lecture#16
mmcky wants to merge 7 commits intomainfrom
translation-sync-2026-04-08T00-26-49-pr-513

mmcky commented Apr 8, 2026

Uh oh!

github-actions bot commented Apr 8, 2026

Vectorization with `vmap` section, paragraph 1: '它能自动将为单个输入编写的函数向量化' contains an extra space before '向量化'. Should be '它能自动将为单个输入编写的函数向量化'.

Overall recommendations section: '对于向量化操作' and '对于顺序操作' use '操作' while earlier sections consistently use '运算' (e.g., '向量化运算', '顺序运算' in headings). Should use '运算' for consistency: '对于向量化运算' and '对于顺序运算'.

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Copilot AI Apr 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

-  name: python3
+  name: python3
+translation:
+  title: Adventures in Autodiff
+  headings:
+    - - Overview
+      - 概述

	JAX 的另一个强大变换是 `jax.vmap`，它能自动将为单个输入编写的函数向量化，使其可以在批量数据上运行。
	JAX 的另一个强大变换是 `jax.vmap`，它能自动将为单个输入编写的函数向量化，使其可以在批量数据上运行。


		ax.text(3, -0.5, "⋮", ha='center', va='center', fontsize=14)

		ax.set_title("PRNG Key Splitting Tree", fontsize=13, pad=10)

	ax.set_title("PRNG Key Splitting Tree", fontsize=13, pad=10)
	ax.set_title("PRNG 密钥拆分树（PRNG Key Splitting Tree）", fontsize=13, pad=10)

Conversation

mmcky commented Apr 8, 2026

Automated Translation Sync

Source PR

Files Added

Files Updated

Details

Uh oh!

github-actions bot commented Apr 8, 2026

✅ Translation Quality Review

📝 Translation Quality

Vectorization with vmap section, paragraph 1: '它能自动将为单个输入编写的函数 向量化' contains an extra space before '向量化'. Should be '它能自动将为单个输入编写的函数向量化'.

Overall recommendations section: '对于向量化操作' and '对于顺序操作' use '操作' while earlier sections consistently use '运算' (e.g., '向量化运算', '顺序运算' in headings). Should use '运算' for consistency: '对于向量化运算' and '对于顺序运算'.

🔍 Diff Quality

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 8, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Vectorization with `vmap` section, paragraph 1: '它能自动将为单个输入编写的函数向量化' contains an extra space before '向量化'. Should be '它能自动将为单个输入编写的函数向量化'.