Unsupervised and controllable synthesizing for imbalanced energy dataset based on AC-InfoGAN

Zhenghao Zhou, Yiyan Li, Runlong Liu, Xiaoyuan Xu and Zheng Yan

Applied Energy, 2025, vol. 393, issue C, No S0306261925008372

Abstract: Generating synthetic data has become a popular alternative solution to deal with the difficulties in accessing and sharing field measurement data in power systems. However, to make the generation results controllable, existing methods (e.g., Conditional Generative Adversarial Nets, cGAN) require labeled dataset to train the model, which is demanding in practice because many field measurement data lack descriptive labels. Meanwhile, real-world datasets are naturally imbalanced, causing bias in neural network training. In this paper, we introduce the Adaptive and Contrastive Information Maximizing Generative Adversarial Nets (AC-InfoGAN) to achieve controllable synthesizing for the unlabeled and imbalanced energy dataset. Features with physical meanings can be automatically extracted by maximizing the mutual information between the input latent code and the classifier output. Then the extracted features are used to control the generation results similar to a vanilla cGAN framework. We employ the Gumbel-Softmax distribution and frequency-based contrastive learning techniques to dynamically adapt to the imbalanced dataset to avoid the model training bias. Meanwhile, frequency-domain neural network modules are introduced to the AC-InfoGAN framework to enhance the model performances. Case study is based on the unlabeled and imbalanced energy datasets of power load and renewable energy output. Results demonstrate that AC-InfoGAN can extract both discrete and continuous features with certain physical meanings, as well as generating realistic synthetic energy data that satisfy given features

Keywords: Information maximizing generative adversarial nets; Unsupervised learning; Unlabeled and imbalanced dataset; Synthetic data generation; Feature extraction (search for similar items in EconPapers)
Date: 2025
References: Add references at CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261925008372
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:393:y:2025:i:c:s0306261925008372

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2025.126107

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().