Error Aware Distribution Prediction for Lightweight Implicit Neural Representations

2026-07-10

AI Summary

This work addresses the inherent prediction errors in implicit neural representations (INRs), which act as lossy approximators, and the limitations of existing uncertainty estimation methods that are either computationally expensive or reliant on strong distributional assumptions. The authors propose reframing regression-based INRs as classification tasks by discretizing continuous targets into intervals. This approach enables lightweight and flexible uncertainty modeling without requiring predefined parametric distributions—such as Gaussian—or complex computations. By doing so, the method effectively captures complex error structures, including multimodal uncertainties, while maintaining high reconstruction fidelity and significantly enhancing the model’s awareness of its own prediction errors.

Abstract

Implicit neural representations (INRs) offer compact encoding of volumes, but as lossy approximators, inevitably have prediction errors. We consider INRs that can simultaneously encode relative error scales by predicting distributions using tools from uncertainty estimation. Typically, uncertainty estimation relies on computationally expensive approaches or on predefined parametric assumptions about the predictive distribution (e.g., Gaussian). In this study, we propose a lightweight method that reformulates regression-based INR training as a classification task by discretizing continuous targets into bins, enabling flexible distribution modeling to capture complex multimodal behaviors. We analyze the trade-off between regression and classification for INR training and demonstrate that the classification setting tends to achieve high reconstruction quality and competitive error awareness through uncertainty estimation, compared to regression-based approaches.

Problem

• Implicit Neural Representations
• Uncertainty Estimation
• Error Awareness
• Distribution Prediction
• Lightweight Modeling

Innovation

• Implicit Neural Representations
• Uncertainty Estimation
• Distribution Prediction
• Lightweight Modeling
• Classification-based Regression