fixes 2

report_thesis/src/sections/problem_definition.tex

@@ -11,7 +11,7 @@ \section{Problem Definition}\label{sec:problem_definition}
 Lastly, \citet{sunMachineLearningTransfer2021} demonstrated the efficacy of transfer learning in overcoming matrix effects and improving model robustness for rock classification on Mars. Their findings suggest that similar improvements can be achieved in oxide quantification by leveraging knowledge from related domains.
 Studies such as these provide a strong foundation for our assumption that addressing the identified challenges will lead to significant improvements in the accuracy and robustness of predicting elemental concentrations from \gls{libs} data.
 
-\subsection{Quantification based on \gls{libs} data}\label{subsec:quantification}
+\subsection{Quantification Based on LIBS Data}\label{subsec:quantification}
 \gls{libs} spectral data provides intensity readings across a spectrum of wavelengths in the form of Clean, Calibrated Spectra \cite{andersonImprovedAccuracyQuantitative2017}, as described by \citet{wiensPreflightCalibrationInitial2013}.
 The wavelength intensities are quantified in units of photon/shot/mm\textsuperscript{2}/sr/nm.