Improving Long-Horizon Forecasts with Expectation-Biased LSTM Networks

TL;DR

This paper introduces expectation-biasing techniques for LSTM networks to enhance long-horizon forecasting accuracy, addressing the decay in performance over extended future predictions in neuroscience and energy datasets.

Contribution

It proposes novel expectation-biasing methods and architectures that significantly improve long-term forecast accuracy over standard LSTM models.

Findings

Expectation-biasing improves long-horizon forecast accuracy.

Proposed architectures outperform standard LSTM models.

Effective on neuroscience and energy datasets.

Abstract

State-of-the-art forecasting methods using Recurrent Neural Net- works (RNN) based on Long-Short Term Memory (LSTM) cells have shown exceptional performance targeting short-horizon forecasts, e.g given a set of predictor features, forecast a target value for the next few time steps in the future. However, in many applica- tions, the performance of these methods decays as the forecasting horizon extends beyond these few time steps. This paper aims to explore the challenges of long-horizon forecasting using LSTM networks. Here, we illustrate the long-horizon forecasting problem in datasets from neuroscience and energy supply management. We then propose expectation-biasing, an approach motivated by the literature of Dynamic Belief Networks, as a solution to improve long-horizon forecasting using LSTMs. We propose two LSTM ar- chitectures along with two methods for expectation biasing that significantly outperforms standard practice.