A Semantic-Loss Function Modeling Framework With Task-Oriented Machine Learning Perspectives

TL;DR

This work addresses semantic loss in Earth Observation communications under bandwidth constraints by proposing a data-fitting framework that links EO objectives to two loss components: source coding loss and transmission loss. It develops both a 1D Exp-2 fitting model and a general multi-parameter 2D model to relate EO task accuracy to data quality $q$ and channel condition ratio $s$, and validates them using EuroSAT imagery processed through a SatCom pipeline with impairments across four ML backbones (EfficientViT, MobileViT, ResNet50-DINO, ResNet8-KD). The empirical results show accurate fits (MAPE < 0.25% with sufficient model complexity) and demonstrate the framework’s applicability for designing semantic-aware, bandwidth-efficient EO communication strategies. Practically, the framework supports more reliable EO operations in bandwidth-limited scenarios and provides a foundation for optimizing data exchange and resource management in satellite-based EO missions, with potential extension to other EO tasks and channels.

Abstract

The integration of machine learning (ML) has significantly enhanced the capabilities of Earth Observation (EO) systems by enabling the extraction of actionable insights from complex datasets. However, the performance of data-driven EO applications is heavily influenced by the data collection and transmission processes, where limited satellite bandwidth and latency constraints can hinder the full transmission of original data to the receivers. To address this issue, adopting the concepts of Semantic Communication (SC) offers a promising solution by prioritizing the transmission of essential data semantics over raw information. Implementing SC for EO systems requires a thorough understanding of the impact of data processing and communication channel conditions on semantic loss at the processing center. This work proposes a novel data-fitting framework to empirically model the semantic loss using real-world EO datasets and domain-specific insights. The framework quantifies two primary types of semantic loss: (1) source coding loss, assessed via a data quality indicator measuring the impact of processing on raw source data, and (2) transmission loss, evaluated by comparing practical transmission performance against the Shannon limit. Semantic losses are estimated by evaluating the accuracy of EO applications using four task-oriented ML models, EfficientViT, MobileViT, ResNet50-DINO, and ResNet8-KD, on lossy image datasets under varying channel conditions and compression ratios. These results underpin a framework for efficient semantic-loss modeling in bandwidth-constrained EO scenarios, enabling more reliable and effective operations.

Figures (11)

• Figure 1: Block diagram of SatCom-integrated data acquisition
• Figure 2: The loss resulting from source information reduction.
• Figure 3: The loss resulting from imperfect wireless transmission.
• Figure 4: Curve fitting model with different models.
• Figure 5: Normalized data size using 'Exp-2' fitting model.