Improving Sample Efficiency of Deep Learning Models in Electricity Market

TL;DR

This paper introduces Knowledge-Augmented Training (KAT), a framework that enhances deep learning sample efficiency in electricity markets by integrating domain knowledge and synthetic data generation to prevent overfitting.

Contribution

The paper proposes a novel KAT framework that combines analytical models with deep learning, including a new data augmentation technique, to improve sample efficiency in electricity market applications.

Findings

KAT outperforms existing methods in user modeling and price forecasting.

Synthetic data generation enhances model robustness.

The approach effectively reduces overfitting in data-scarce scenarios.

Abstract

The superior performance of deep learning relies heavily on a large collection of sample data, but the data insufficiency problem turns out to be relatively common in global electricity markets. How to prevent overfitting in this case becomes a fundamental challenge when training deep learning models in different market applications. With this in mind, we propose a general framework, namely Knowledge-Augmented Training (KAT), to improve the sample efficiency, and the main idea is to incorporate domain knowledge into the training procedures of deep learning models. Specifically, we propose a novel data augmentation technique to generate some synthetic data, which are later processed by an improved training strategy. This KAT methodology follows and realizes the idea of combining analytical and deep learning models together. Modern learning theories demonstrate the effectiveness of our method in terms of effective prediction error feedbacks, a reliable loss function, and rich gradient noises. At last, we study two popular applications in detail: user modeling and probabilistic price forecasting. The proposed method outperforms other competitors in all numerical tests, and the underlying reasons are explained by further statistical and visualization results.