On What We Can Learn from Low-Resolution Data

TL;DR

This paper investigates the informativeness of low-resolution data in AI systems, providing theoretical bounds and empirical evidence that low-resolution data can enhance model performance when high-resolution data is limited.

Contribution

It offers a theoretical framework based on Kullback-Leibler divergence to quantify low-resolution data's impact and empirically shows its benefits in training models.

Findings

Adding low-resolution data improves model performance with scarce high-resolution data.

Theoretical bounds relate information loss to data resolution.

Empirical results confirm the value of low-resolution data in vision models.

Abstract

Artificial intelligence systems typically rely on large, centrally collected datasets, a premise that does not hold in many real-world domains such as healthcare and public institutions. In these settings, data sharing is often constrained by storage, privacy, or resource limitations. For example, small wearable devices may lack the bandwidth or energy capacity needed to store and transmit high-resolution data, leading to aggregation during data collection and thus a loss of information. As a result, datasets collected from different sources may consist of a mixture of high- and low-resolution samples. Despite the prevalence of this setting, it remains unclear how informative low-resolution data is when models are ultimately evaluated on high-resolution inputs. We provide a theoretical analysis based on the Kullback-Leibler divergence that characterises how the influence of a datapoint changes with resolution, and derive bounds that relate the relative contribution of high- and low-resolution observations to the information lost under downsampling. To support this analysis, we empirically demonstrate, using both a vision transformer and a convolutional neural network, that adding low-resolution data to the training set consistently improves performance when high-resolution data is scarce.