Memoir Thesis

\documentclass[11pt,a4paper,twoside,openright]{memoir}
\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage{graphicx}
\usepackage{amsmath,amssymb}
\usepackage{booktabs}
\usepackage[backend=biber,style=numeric]{biblatex}
\usepackage[hidelinks]{hyperref}

\addbibresource{references.bib}

\setstocksize{297mm}{210mm}
\settrimmedsize{297mm}{210mm}{*}
\setlrmarginsandblock{3cm}{2.5cm}{*}
\setulmarginsandblock{3cm}{3cm}{*}
\checkandfixthelayout

\chapterstyle{thatcher}
\renewcommand*{\chapnumfont}{\normalfont\Large\scshape}
\renewcommand*{\chaptitlefont}{\normalfont\Huge\bfseries}

\title{A Memoir-Based Dissertation Template}
\author{First Last}
\date{\today}

\begin{document}

\frontmatter
\begin{titlingpage}
  \centering
  \vspace*{3cm}
  {\Huge\bfseries A Memoir-Based Dissertation Template}\\[2cm]
  {\LARGE First Last}\\[3cm]
  A thesis submitted in fulfillment of the requirements for the degree of\\
  Doctor of Philosophy\\[1cm]
  University of Example\\
  School of Example Studies\\[2cm]
  \today
\end{titlingpage}

\chapter*{Abstract}
This work explores a set of methods for robust estimation under distribution shift.
Our contributions include a new estimator, theoretical guarantees, and empirical validation.

\tableofcontents*

\mainmatter

\chapter{Introduction}
\section{Context}
Distribution shift is ubiquitous in deployed ML systems.
\section{Outline}
Chapter 2 reviews the field; Chapter 3 introduces our estimator.

\chapter{Preliminaries}
Let $P_0$ and $P_1$ denote the source and target distributions.

\chapter{Proposed Method}
We minimize a weighted empirical risk with a KL penalty.
\begin{equation}
  \hat\theta = \arg\min_\theta \frac{1}{n}\sum_{i=1}^n w_i \ell(\theta; z_i)
             + \lambda \mathrm{KL}(w \| \mathbf{1}).
\end{equation}

\chapter{Experiments}
We evaluate on CIFAR-10-C, ImageNet-R, and DomainNet.

\chapter{Conclusion}
We have shown clear improvements over strong baselines.

\backmatter
\printbibliography
\end{document}