Task-Conditioned Synthetic Data Generation for Improving Machine Learning Performance in Agricultural Prediction Tasks

Jul 16, 2026

Upload by Skyler

Machine Learning models estimate agricultural variables, yet limited or incomplete reference data can restrict prediction quality because performance depends strongly on training data quantity and representativeness. Task-Conditioned Synthetic Data Generation (TCSDG) addresses this by producing realistic synthetic samples that preserve key properties of the original dataset. TCSDG combines a Bayesian Network generator with a transformer-based tabular foundation model (TabICL) using teacher–student knowledge transfer and in-context learning. Evaluations on crop yield prediction and crop type classification across twelve sites show consistent gains over multiple settings and improved performance in 89% and 74% of experiments.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages24

Reading Time1 h 0 min

Abstract

Introduction

Agricultural prediction with machine learning

Dependence on training data quality and quantity

Limits of agricultural reference data

arXiv :2607 .0975 1v 1 [ cs .AI] 4 Jul 2026 Task-Conditioned Synthetic Data Generation for Improving Machine Learning Performance in Agricultural Prediction Tasks Hamid Ebrahimy 1,2,* Moritz Lucas 1,2 Martin Atzmueller 1,3 1 Osnabrück University, Joint Lab Artificial Intelligence and Data Science, Osnabrück, Germany 2 Leibniz Institute for Agricultural Engineering and Bioeconomy (ATB), Potsdam, Germany 3 German Research Center for Artificial Intelligence (DFKI), Research Department Cooperative and Autonomous Systems (CAS), Osnabrück, Germany * Corresponding author: [hamid.ebrahimy@uni-osnabrueck.de](hamid.ebrahimy@uni-osnabrueck.de) Abstract Machine Learning (ML) algorithms have been widely used to estimate agricultural variables across diverse contexts. However, because the quantity and quality of training data strongly influence performance of ML algorithms, their use can be constrained by limited or incomplete reference data. Synthetic Data Generation (SDG) offers a practical approach to address this issue by producing artificial but realistic samples that preserve key characteristics of the original data. Building on teacher–student knowledge transfer and incontext learning for tabular data, this study proposes a Task-Conditioned SDG (TCSDG) algorithm that pairs a Bayesian Network generator with a transformer-based tabular foundation model (TabICL) . The proposed algorithm was evaluated on two agricultural prediction tasks: crop yield prediction and crop type classification. Six benchmark SDG algorithms were also utilized to compare their performance with that of TCSDG. Across twelve study sites, two training-data fractions, four multiplication ratios, and three predictive ML algorithms, augmenting the original data with TCSDG-generated synthetic data improved ML performance in 89% of the crop type classification experiments and 74% of the crop yield prediction experiments. TCSDG also substantially outperformed benchmark SDG algorithms and was the only method to consistently improve ML performance across both tasks at the aggregate level. The study demonstrates that carefully designed and processed synthetic data can improve ML performance in precision-agriculture applications. TCSDG offers a practical and extensible framework for generating synthetic data that supports downstream ML agricultural prediction. The full implementation of TCSDG is publicly available as opensource at [https://github.com/HamidEbrahimy/TCSDG](https://github.com/HamidEbrahimy/TCSDG). 1. Introduction Agricultural variables are critical indicators that have been used in various applications including food security assessment, agricultural optimization, economic planning, and climate change impact evaluation (Bouguettaya et al., 2022; Gallego et al., 2010; Van Klompenburg et al., 2020) . With the population growth and the increasing impact of climate change on agriculture, continuous monitoring and evalua- tions of agricultural variables is vital and remains critical for achieving long-term agricultural sustainability and food security (Becker-Reshef et al., 2023; Challinor et al., 2014; Kang et al., 2009) . Accordingly, in view of the importance of agricultural variables across a wide range of applications, substantial advances in data analytics, remote sensing technologies, and Machine Learning (ML) algorithms have been leveraged for their assessment (Muruganantham et al., 2022; Pierre Pott et al., 2022) . ML algorithms are widely employed for the prediction of different agricultural variables (Feng et al. , 2019; Rashid et al., 2021; Tariq et al., 2023) . These algorithms are able to process vast amounts of heterogeneous data, including meteorological variables, soil characteristics, crop health indicators, and remote sensing data. In a broad sense, supervised ML models apply advanced computational strategies to analyze relationships between specific agricultural independent variables (e.g. , soil moisture, climate cond

FAQs

Why does synthetic data help agricultural ML prediction tasks?

Synthetic Data Generation creates artificial but realistic samples that preserve key characteristics of the original training data. This helps when reference data are limited, incomplete, or low quality, improving model learning and generalization.

How does TCSDG generate task-conditioned synthetic data?

TCSDG pairs a Bayesian Network generator with a transformer-based tabular foundation model (TabICL). It uses teacher–student knowledge transfer and in-context learning for tabular data to condition generation on the task.

What were TCSDG’s reported results on agricultural tasks?

On crop type classification and crop yield prediction, TCSDG improved ML performance in 89% and 74% of experiments, respectively. It also outperformed six benchmark SDG algorithms and was the only method consistently improving both tasks at the aggregate level.

Jul 16, 2026

Upload by Skyler