Making Good on LSTMs' Unfulfilled Promise

TL;DR

This paper demonstrates that memory-augmented continual learning approaches, especially those based on feed-forward neural networks, outperform traditional LSTMs in real-world financial time-series tasks, offering better explainability and robustness.

Contribution

It introduces Continual Learning Augmentation (CLA) as a flexible framework, compares different learners and similarity measures, and highlights the superiority of FFNN-based CL over LSTM-based CL in financial applications.

Findings

FFNN-based CL outperforms LSTM-based CL and simple sliding window models.

Warp-AE similarity approach yields the best performance in memory recall.

Time-series noise impacts similarity measures, with warp-AE being most robust.

Abstract

LSTMs promise much to financial time-series analysis, temporal and cross-sectional inference, but we find that they do not deliver in a real-world financial management task. We examine an alternative called Continual Learning (CL), a memory-augmented approach, which can provide transparent explanations, i.e. which memory did what and when. This work has implications for many financial applications including credit, time-varying fairness in decision making and more. We make three important new observations. Firstly, as well as being more explainable, time-series CL approaches outperform LSTMs as well as a simple sliding window learner using feed-forward neural networks (FFNN). Secondly, we show that CL based on a sliding window learner (FFNN) is more effective than CL based on a sequential learner (LSTM). Thirdly, we examine how real-world, time-series noise impacts several similarity approaches used in CL memory addressing. We provide these insights using an approach called Continual Learning Augmentation (CLA) tested on a complex real-world problem, emerging market equities investment decision making. CLA provides a test-bed as it can be based on different types of time-series learners, allowing testing of LSTM and FFNN learners side by side. CLA is also used to test several distance approaches used in a memory recall-gate: Euclidean distance (ED), dynamic time warping (DTW), auto-encoders (AE) and a novel hybrid approach, warp-AE. We find that ED under-performs DTW and AE but warp-AE shows the best overall performance in a real-world financial task.