PDF bookmarks for "Machine Learning Yearning (Draft Version) - Andrew Ng" (LaTeX)

Machine Learning Yearning Bookmarks.md

PDF bookmarks for "Machine Learning Yearning (Draft Version) - Andrew Ng" (LaTeX)

This gist contains out.tex, a tex file that adds a PDF outline ("bookmarks") to the freely available pdf file of the book

Machine Learning Yearning: Technical Strategy for AI Engineers In the Era of Deep Learning (Draft Version), by Andrew Ng

available from https://www.deeplearning.ai/machine-learning-yearning/

The bookmarks allow to navigate the contents of the book while reading it on a screen.

Usage

• complete form at https://www.deeplearning.ai/machine-learning-yearning/
• download Ng-MLY01-13.pdf from the email url (this code may not work for newer versions of the file)
• download out.tex into the same folder as Ng-MLY01-13.pdf
• compile as pdflatex out.tex
• rename the resulting output file out.pdf to e.g. Machine Learning Yearning - Andrew Ng.pdf

Credit for gist format

Format of this gist copied from: https://gist.github.com/goerz/4c863a2fde1d3357113b95643d0ace16

out.tex

% Usage:
% * input email information at: https://www.deeplearning.ai/machine-learning-yearning/
% * using url from email, download 'Ng-MLY01-13.pdf'
% * store this file as 'out.tex', and compile as 'pdflatex out.tex'
% * rename output file to e.g.
%   'Machine Learning Yearning - Andrew Ng.pdf'

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage{geometry}


\usepackage{pdfpages}
\usepackage[
  pdfpagelabels=true,
  pdftitle={Machine Learning Yearning (Draft Version)},
  pdfauthor={Andrew Ng},
  pdfsubject={Machine Learning},
  pdfkeywords={machine learning},
  unicode=true,
]{hyperref}
\usepackage{bookmark}

\begin{document}

\pagenumbering{arabic}
\setcounter{page}{1}
\includepdf[pages=1-]{Ng-MLY01-13.pdf}

\bookmark[page=1,level=0]{Cover}
\bookmark[page=3,level=0]{Table of Contents}

\bookmark[page=6,level=0]{1 Why Machine Learning Strategy}
\bookmark[page=8,level=0]{2 How to use this book to help your team}
\bookmark[page=9,level=0]{3 Prerequisites and Notation}
\bookmark[page=10,level=0]{4 Scale drives machine learning progress}
\bookmark[page=14,level=0]{5 Your development and test sets}
\bookmark[page=17,level=0]{6 Your dev and test sets should come from the same distribution}
\bookmark[page=19,level=0]{7 How large do the dev/test sets need to be?}
\bookmark[page=20,level=0]{8 Establish a single-number evaluation metric for your team to optimize}
\bookmark[page=22,level=0]{9 Optimizing and satisficing metrics}
\bookmark[page=24,level=0]{10 Having a dev set and metric speeds up iterations}
\bookmark[page=25,level=0]{11 When to change dev/test sets and metrics}
\bookmark[page=27,level=0]{12 Takeaways: Setting up development and test sets}
\bookmark[page=29,level=0]{13 Build your first system quickly, then iterate}
\bookmark[page=30,level=0]{14 Error analysis: Look at dev set examples to evaluate ideas}
\bookmark[page=32,level=0]{15 Evaluating multiple ideas in parallel during error analysis}
\bookmark[page=34,level=0]{16 Cleaning up mislabeled dev and test set examples}
\bookmark[page=36,level=0]{17 If you have a large dev set, split it into two subsets, only one of which you look at}
\bookmark[page=38,level=0]{18 How big should the Eyeball and Blackbox dev sets be?}
\bookmark[page=40,level=0]{19 Takeaways: Basic error analysis}
\bookmark[page=42,level=0]{20 Bias and Variance: The two big sources of error}
\bookmark[page=44,level=0]{21 Examples of Bias and Variance}
\bookmark[page=46,level=0]{22 Comparing to the optimal error rate}
\bookmark[page=49,level=0]{23 Addressing Bias and Variance}
\bookmark[page=50,level=0]{24 Bias vs. Variance tradeoff}
\bookmark[page=51,level=0]{25 Techniques for reducing avoidable bias}
\bookmark[page=52,level=0]{26 Error analysis on the training set}
\bookmark[page=53,level=0]{27 Techniques for reducing variance}
\bookmark[page=56,level=0]{28 Diagnosing bias and variance: Learning curves}
\bookmark[page=59,level=0]{29 Plotting training error}
\bookmark[page=60,level=0]{30 Interpreting learning curves: High bias}
\bookmark[page=62,level=0]{31 Interpreting learning curves: Other cases}
\bookmark[page=63,level=0]{32 Plotting learning curves}
\bookmark[page=66,level=0]{33 Why we compare to human-level performance}
\bookmark[page=68,level=0]{34 How to define human-level performance}
\bookmark[page=69,level=0]{35 Surpassing human-level performance}
\bookmark[page=71,level=0]{36 When you should train and test on different distributions}
\bookmark[page=73,level=0]{37 How to decide whether to use all your data}
\bookmark[page=75,level=0]{38 How to decide whether to include inconsistent data}
\bookmark[page=76,level=0]{39 Weighting data}
\bookmark[page=77,level=0]{40 Generalizing from the training set to the dev set}
\bookmark[page=79,level=0]{41 Identifying Bias, Variance, and Data Mismatch Errors}
\bookmark[page=81,level=0]{42 Addressing data mismatch}
\bookmark[page=82,level=0]{43 Artificial data synthesis}
\bookmark[page=85,level=0]{44 The Optimization Verification test}
\bookmark[page=87,level=0]{45 General form of Optimization Verification test}
\bookmark[page=88,level=0]{46 Reinforcement learning example}
\bookmark[page=91,level=0]{47 The rise of end-to-end learning}
\bookmark[page=93,level=0]{48 More end-to-end learning examples}
\bookmark[page=95,level=0]{49 Pros and cons of end-to-end learning}
\bookmark[page=97,level=0]{50 Choosing pipeline components: Data availability}
\bookmark[page=99,level=0]{51 Choosing pipeline components: Task simplicity}
\bookmark[page=103,level=0]{52 Directly learning rich outputs}
\bookmark[page=106,level=0]{53 Error analysis by parts}
\bookmark[page=109,level=0]{54 Attributing error to one part}
\bookmark[page=111,level=0]{55 General case of error attribution}
\bookmark[page=113,level=0]{56 Error analysis by parts and comparison to human-level performance}
\bookmark[page=115,level=0]{57 Spotting a flawed ML pipeline}
\bookmark[page=118,level=0]{58 Building a superhero team}


\end{document}