MULTI- LAYER 5D OPTICAL DATA STORAGE: MATHEMATICAL MODELING AND DEEP LEARNING- BASED RECONSTRUCTION OF BIREFRINGENT PARAMETERS (726,00 руб.)

Первый автор	Zhang
Авторы	Zhu Q., Zhou R., Lysak T.
Страниц	15

ID	950867
Аннотация	Five- dimensional ( 5D) optical data storage has recently emerged as a promising solution for ultra- high- density and long- term data archiving [ 1] –[ 5] . It encodes information using three spatial coordinates ( x, y, z) , combined with two optical parameters derived from laser- induced nanostructures [ 6] –[ 9] . Specifically, femtosecond laser pulses generate birefringent nanogratingsinsidefusedsilica. Thus information isencodedoneachdatavoxel through its position, the slowaxis orientation, and the magnitude of optical retardance. This five- dimensional encodeing scheme enables extraordinary storage capacity and durability. For example, a single nanostructured quartz disc can theoretically store the order of 102 ∼ 103 terabytes while maintaining stability at temperatures up to 1000◦C [ 1] , making it highly suitable for long- term archival storage. However, practical realization of 5D optical storage remains challenged by noise and interference during data readout [ 10] –[ 13] . This significantly affects the accurate reconstruction of the encoded optical parameters. One major issue is the interlayer crosstalk. As voxel layers are packed closer together to increase storage density, signals from adjacent layersoverlap, reducing decoding accuracy [ 14] . Furthermore, signal attenuation occursas the probe beam travels through multiple layers, with cumulative absorption and scattering diminishing signal quality. Besides, additional degradation also arises from external noise sources, including detector fluctuations, background light, and instability in laser polarization. As a result, these factors lead to a complex, ill- posed inverse problem, where small measurement errors can result in large inaccuracies in retrieving voxel- specific optical parameters. This often forces compromise in encoding strategies, such as limiting the range of birefringent values to ensure decoding reliability. Altogether, robust and accurate parameter recovery from noisy and interfered measurements remains a central challenge for scalable 5D optical memory systems. The core technical problem of reading 5D optical storage can be formulated as the inversion of polarization- resolved intensity data. Its goal is to reconstruct the slow- axis direction ψ and retardance magnitudes Δ for each voxel. This

MULTI- LAYER 5D OPTICAL DATA STORAGE: MATHEMATICAL MODELING AND DEEP LEARNING- BASED RECONSTRUCTION OF BIREFRINGENT PARAMETERS / Y. Zhang [и др.] // Журнал вычислительной математики и математической физики (РАН) .— 2026 .— №1 .— С. 120-134 .— URL: https://rucont.ru/efd/950867 (дата обращения: 12.04.2026)