slds-lmu / teaching_devops_issues

Meta-repo for teaching infrastructure issues
0 stars 0 forks source link

Unified lecture start slide (LaTeX macro) #10

Closed jemus42 closed 4 months ago

jemus42 commented 5 months ago

Looking at this and the start components of each slides, e.g. from https://github.com/slds-lmu/lecture_i2ml/blob/master/slides/cart/slides-cart-computationalaspects.tex

\documentclass[11pt,compress,t,notes=noshow, xcolor=table]{beamer}
\input{../../style/preamble}
\input{../../latex-math/ml-trees.tex}

\newcommand{\titlefigure}{figure/categoryplot-binarysmall.pdf}
\newcommand{\learninggoals}{
\item Know how monotone feature transformations affect the tree
\item Understand how categorical features can be treated effectively while growing a CART
\item Understand how missing values can be treated in a CART}

\title{Introduction to Machine Learning}
% \author{Bernd Bischl, Christoph Molnar, Daniel Schalk, Fabian Scheipl}
\institute{\href{https://compstat-lmu.github.io/lecture_i2ml/}{compstat-lmu.github.io/lecture\_i2ml}}
\date{}

\begin{document}
\lecturechapter{CART: \\
Computational Aspects of Finding Splits}
\lecture{Introduction to Machine Learning}

\sloppy

\begin{vbframe}{Monotone feature transformations}

[...]

It's not clear how to unify the \titlefigure, \learninggoals and later \lecture into a single command, especially with \title having to be placed before \begin{document} and things like \lecturechapter needing to be placed after the \begin{document}.

jemus42 commented 5 months ago

The current approach is now a vast reduction of complexity that just relies on a single macro, see https://github.com/slds-lmu/lecture_i2ml/pull/1211

It's not perfect and some flexibility might be lost, but to the best of my knowledge it works and appears to be functionally equivalent to the previous implementation, and at least in i2ml I was still able to successfully render the combined slides with adapted slide numbers and everything.

jemus42 commented 4 months ago

This is integrated in the service files now and rolled out in i2ml/sl, with optim on the way and advml on the horizon