Adaptive Multi-Gradient Guidance with Conflict Resolution for Limited-Sample Regression
Open Access
- 21 July 2025
- journal article
- research article
- Published by MDPI AG in Information
- Vol. 16 (7) , 619
Abstract
Recent studies report that gradient guidance extracted from a single-reference model can improve Limited-Sample regression. However, one reference model may not capture all relevant characteristics of the target function, which can restrict the capacity of the learner. To address this issue, we introduce the Multi-Gradient Guided Network (MGGN), an extension of single-gradient guidance that combines gradients from several reference models. The gradients are merged through an adaptive weighting scheme, and an orthogonal-projection step is applied to reduce potential conflicts between them. Experiments on sine regression are used to evaluate the method. The results indicate that MGGN achieves higher predictive accuracy and improved stability than existing single-gradient guidance and meta-learning baselines, benefiting from the complementary information provided by multiple reference models.Keywords
Funding Information
- Fujian Provincial Natural Science Foundation of China (2021J01124, 2024I1001)
- Fujian Provincial Science and Technology Plan Project of China (2021J01124, 2024I1001)
This publication has 25 references indexed in Scilit:
- Wi-Fi fingerprint based indoor localization using few shot regression2024
- An Automated Few-Shot Learning for Time-Series Forecasting in Smart Grid Under Data ScarcityIEEE Transactions on Artificial Intelligence, 2024
- Few-shot regression with differentiable reference modelInformation Sciences, 2023
- SSL-ProtoNet: Self-supervised Learning Prototypical Networks for few-shot learningExpert Systems with Applications, 2023
- Learning to Learn Task-Adaptive Hyperparameters for Few-Shot LearningIEEE Transactions on Pattern Analysis and Machine Intelligence, 2023
- An Adversarial Meta-Training Framework for Cross-Domain Few-Shot LearningIEEE Transactions on Multimedia, 2022
- Contrastive Decoder Generator for Few-Shot Learning in Product Quality PredictionIEEE Transactions on Industrial Informatics, 2022
- Few-shot imbalanced classification based on data augmentationMultimedia Systems, 2021
- Generalizing from a Few ExamplesACM Computing Surveys, 2020
- Small sample adjustments for robust variance estimation with meta-regression.Psychological Methods, 2015