Multi-Layered Gradient Boosting Decision Trees
This work addresses the problem of representation learning in non-differentiable models like GBDTs, which are widely used for tabular data, by introducing a novel multi-layered approach, though it appears incremental in extending GBDTs with deep learning concepts.
The paper tackles the challenge of enabling gradient boosting decision trees (GBDTs) to learn hierarchical representations, similar to deep neural networks, by proposing multi-layered GBDT forests (mGBDTs) that use stacked regression GBDTs and target propagation for joint training, resulting in improved performance and representation learning ability as confirmed by experiments and visualizations.
Multi-layered representation is believed to be the key ingredient of deep neural networks especially in cognitive tasks like computer vision. While non-differentiable models such as gradient boosting decision trees (GBDTs) are the dominant methods for modeling discrete or tabular data, they are hard to incorporate with such representation learning ability. In this work, we propose the multi-layered GBDT forest (mGBDTs), with an explicit emphasis on exploring the ability to learn hierarchical representations by stacking several layers of regression GBDTs as its building block. The model can be jointly trained by a variant of target propagation across layers, without the need to derive back-propagation nor differentiability. Experiments and visualizations confirmed the effectiveness of the model in terms of performance and representation learning ability.