mirror of
https://github.com/zoo-apps/papers.git
synced 2026-10-30 14:56:54 +00:00
30 lines
1.2 KiB
TeX
30 lines
1.2 KiB
TeX
\section{Ablation Studies}
|
|
|
|
\label{sec:ablation}
|
|
|
|
We conducted ablation experiments on the three primary system components to quantify their individual contributions:
|
|
|
|
\begin{table}[H]
|
|
\centering
|
|
\caption{Ablation study results: 12-month retention and field transfer scores.}
|
|
\label{tab:ablation}
|
|
\begin{tabular}{@{}lcc@{}}
|
|
\toprule
|
|
\textbf{Configuration} & \textbf{12-Month Retention} & \textbf{Field Transfer} \\
|
|
\midrule
|
|
Full System & 68.9 & 78.7 \\
|
|
$-$ Ecological Grounding & 55.4 ($-$13.5) & 72.4 ($-$6.3) \\
|
|
$-$ Adaptive Sequencing & 59.2 ($-$9.7) & 70.1 ($-$8.6) \\
|
|
$-$ Multi-Modal Assessment & 64.7 ($-$4.2) & 73.8 ($-$4.9) \\
|
|
$-$ All (Static Baseline) & 44.3 ($-$24.6) & 66.6 ($-$12.1) \\
|
|
\bottomrule
|
|
\end{tabular}
|
|
\end{table}
|
|
|
|
Ecological grounding has the largest individual effect on retention, while adaptive sequencing contributes most to field transfer.
|
|
This is consistent with the hypothesis that sensor-triggered reviews strengthen memory consolidation, while personalized learning paths develop deeper procedural competence.
|
|
|
|
% ============================================================
|
|
% 9. Discussion
|
|
% ============================================================
|