Decision Trees That Remember: Gradient-Based Learning of Recurrent Decision Trees with Memory

TL;DR

ReMeDe Trees are a new type of recurrent decision tree model that incorporates memory and learns long-term dependencies in sequential data using gradient-based optimization.

Contribution

Introduction of ReMeDe Trees, a recurrent decision tree architecture with internal memory, capable of capturing complex temporal dependencies directly from sequential data.

Findings

Effective on synthetic benchmarks for sequential data

Learns long-term dependencies via gradient descent

Integrates memory mechanism similar to RNNs

Abstract

Neural architectures such as Recurrent Neural Networks (RNNs), Transformers, and State-Space Models have shown great success in handling sequential data by learning temporal dependencies. Decision Trees (DTs), on the other hand, remain a widely used class of models for structured tabular data but are typically not designed to capture sequential patterns directly. Instead, DT-based approaches for time-series data often rely on feature engineering, such as manually incorporating lag features, which can be suboptimal for capturing complex temporal dependencies. To address this limitation, we introduce ReMeDe Trees, a novel recurrent DT architecture that integrates an internal memory mechanism, similar to RNNs, to learn long-term dependencies in sequential data. Our model learns hard, axis-aligned decision rules for both output generation and state updates, optimizing them efficiently via gradient descent. We provide a proof-of-concept study on synthetic benchmarks to demonstrate the effectiveness of our approach.