Расширил описание в ml для этой статьи

report.tex

@@ -158,7 +158,7 @@
 
     The authors of~\cite{tabular} developed a machine learning model to predict cisplatin sensitivity based on gene expression changes induced by cisplatin treatment. They combined gene expression data from sensitive ovarian cancer cell lines and patients with specific signaling alterations to identify a gene signature. Using this signature, they trained TabNet, an interpretable deep learning algorithm for tabular data, to perform binary classification of sensitivity to cisplatin. Also several other machine learning algorithms, including Ridge, LASSO, Elastic Net, Nu-Support Vector Classification (Nu-SVC), XGBoost, and Random Forest, were applied to the same task for comparission with TabNet.
 
-    Same as in the~\cite{heterogeneity}, the authors of~\cite{deep} used algorithms from the specialized software called Acapella (developed by PerkinElmer~\cite{PerkinElmer}) to extract 624 quantitative image features from cellular images. This information was fed into a deep learning model that identified a continuous 27-dimension space describing all of the observed cell morphologies. After that the random forest classifier was trained on populations of cells labeled as either drug sensitive or drug resistant.
+    Same as in the~\cite{heterogeneity}, the authors of~\cite{deep} used algorithms from the specialized software called Acapella (developed by PerkinElmer~\cite{PerkinElmer}) to extract 624 quantitative image features from cellular images. These quantified features were then analyzed using deep learning in order to reduce the dimensionality of the data. The authors ran 46 separate deep learning models, all with the same input data but with different numbers of output nodes in order to reduce the dimensionality of the data to greater and lesser extents. Then they scored these 46 different models based on their ability to recreate the input data. After that they selected the model with the lowest number of dimensions, in this case 27, that reached reconstruction error plateau. This deep learning model identified a continuous 27-dimension space describing all of the observed cell morphologies. Upon this model a random forest classifier was trained on populations of cells labeled as either drug sensitive or drug resistant.
 
 
     \section{Datasets}