CIF: Small: Towards Knowledge-Based Topologically Informed Error Detection and Correction

Emerging extended reality technologies, including augmented reality, mixed reality, and virtual reality, demand high-speed low-latency communication to support interactive immersive applications such as three-dimensional (3D) telemedicine and video conferencing. However, achieving both high speed and low latency can lead to a high bit error rate, which in turn reduces communication reliability. Thus, it is essential to develop a new error detection and correction coding solution that can tolerate communication errors in such scenarios. In communication systems, bit errors exhibit varying levels of semantic significance. Errors with low significance can be tolerated, while those with high significance must be corrected. This project will explore topological information from source data for error detection and correction in communication systems, which can ensure data fidelity at the topology level rather than the bit level, aiming to detect the significance of errors in the received data and correct them by recovering the original topological information. This approach offers new insights into communication errors and can potentially enable the development of novel codes for next-generation communication systems. In addition, this project will provide research opportunities for both graduate and undergraduate students and will leverage outreach activities at the University of Nebraska-Lincoln to share research outcomes with K-12 students and teachers using specially developed educational modules.

In this project, topological data analysis, particularly persistent homology and persistence diagrams, will be used to extract and encode topological information from source data. The topological information captures high-level data relations, which can be more compact and semantically meaningful than bit-level relations. This project will study the feasibility and evaluate the performance of topological error detection and correction using point cloud data as a primary data modality. A foundational relation between bit errors and topological errors will be established, aiming to guide the development of topological error detection algorithms using uncompressed persistence diagrams to detect and evaluate the significance of errors. In addition, topological error correction algorithms will be designed by optimizing persistent homology functions to correct significant errors in the received data.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Award Number: 2516013
Principal Investigator: Hongzhi Guo
Funds Obligated: $125,000
State: NE