Applying Artificial neural network into kaggle’s house prices data set gave bad predicted values

I am trying to solve the kaggle’s house prices using neural network. I’ve already made it with ensembling several models (XGBoost, GradientBooster and Ridge) and I’ve got a great score ranking me between the top 25%. I imagined that by adding a new model to the ensembled models like ANN would increase prediction accuracy, so…

How to handle set-like size agnostic input format

Let’s set up some hypothetical simplified scenario: Each instance $i$ of my imaginary dataset $D=\{i_{1}, \ldots, i_{MAX}\}$ has different number $k_{i}$ of $n$-dimensional vectors as input into my neural network. Each of them will be transformed with $m \times n$ matrix $M$ (so, matrices with same parameters) and acted point-wise with some non-linearity $\sigma_{1}$. Now…