PRISM: Position-encoded Regressive Inverse Spectral Model for Multilayer Thin-Film Design

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wang, Runtian, Xue, Renhao, Chen, Baige, Wu, Hao
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913171889979392
author Wang, Runtian
Xue, Renhao
Chen, Baige
Wu, Hao
author_facet Wang, Runtian
Xue, Renhao
Chen, Baige
Wu, Hao
contents The inverse problem of multilayer thin-film optical coatings design represents a complex combinatorial-continuous optimization challenge. We present PRISM (Position-encoded Regressive Inverse Spectral Model), a unified decoder-only autoregressive transformer that streamlines this process by jointly predicting discrete material selection and continuous thickness regression within a single backbone. PRISM introduces two primary architectural innovations: (1) spectrum prefix conditioning, which utilizes standard prefix tokens for in-context target injection, and (2) cumulative-depth Rotary Position Embeddings, which encode continuous thickness directly into the positional representation to preserve the physical spatial relationships of the stack. Our benchmarks demonstrate that a PRISM-13M model reduces MAE by over 50\% compared to other transformer baselines while utilizing only one-fifth of the parameters. Furthermore, a 44M-parameter variant achieves state-of-the-art performance (MAE = 0.010) on our in-distribution validation benchmark and operates significantly faster than simulated annealing, offering a highly efficient alternative to classical optimization methods.
format Preprint
id arxiv_https___arxiv_org_abs_2605_26502
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle PRISM: Position-encoded Regressive Inverse Spectral Model for Multilayer Thin-Film Design
Wang, Runtian
Xue, Renhao
Chen, Baige
Wu, Hao
Machine Learning
Optics
The inverse problem of multilayer thin-film optical coatings design represents a complex combinatorial-continuous optimization challenge. We present PRISM (Position-encoded Regressive Inverse Spectral Model), a unified decoder-only autoregressive transformer that streamlines this process by jointly predicting discrete material selection and continuous thickness regression within a single backbone. PRISM introduces two primary architectural innovations: (1) spectrum prefix conditioning, which utilizes standard prefix tokens for in-context target injection, and (2) cumulative-depth Rotary Position Embeddings, which encode continuous thickness directly into the positional representation to preserve the physical spatial relationships of the stack. Our benchmarks demonstrate that a PRISM-13M model reduces MAE by over 50\% compared to other transformer baselines while utilizing only one-fifth of the parameters. Furthermore, a 44M-parameter variant achieves state-of-the-art performance (MAE = 0.010) on our in-distribution validation benchmark and operates significantly faster than simulated annealing, offering a highly efficient alternative to classical optimization methods.
title PRISM: Position-encoded Regressive Inverse Spectral Model for Multilayer Thin-Film Design
topic Machine Learning
Optics
url https://arxiv.org/abs/2605.26502