Spectra-Scope : A toolkit for automated and interpretable characterization of material properties from spectral data

Amalya C. Johnson, Chris Fajardo, Leena Sansguiri, Weike Ye, Steven B. Torrisi · 2026

Read Spectra-Scope : A toolkit for automated and interpretable characterization of material properties from spectral data by Amalya C. Johnson, Chris Fajardo, Leena Sansguiri, Weike Ye, Steven B. Torrisi (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-06011.txt.

Introduction

Johnson and team present Spectra-Scope, an AutoML frame for matter traits from spectra reading. Spectra reading is central to matter study. Many methods challenge clear guided learning from non-line signal-response links. The software, in Python with a no-code web app, holds cleaning, pulling, training, and picking tools. Users can train simple, clear models on changes with modest means. The work outlines Spectra-Scope methods and power across sets. Spectra-Scope matches similar published scores, and clarity helps explain model ways and real workings behind spectral traits.

Catalog record

CREATORAmalya C. Johnson, Chris Fajardo, Leena Sansguiri, Weike Ye, Steven B. Torrisi
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37677
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.06011

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Spectra-Scope : A toolkit for automated and interpretable characterization of material properties from spectral data Amalya C. Johnson, Chris Fajardo, Leena Sansguiri, Weike Ye, Steven B. Torrisi Abstract Spectroscopy is a central pillar of materials characterization, providing useful information on properties like structure, composition, or excited state dynamics of a system. However, many spectroscopic techniques present challenges in development of interpretable, performant, and reliable supervised learning models due to the wide range of possible nonlinear correlations that can exist between the signal and the response variable (target) of interest. Here, we present Spectra-Scope, an open-source AutoML framework for automatic characterization of material properties from spectroscopy data using interpretable machine learning (ML) models. The software is implemented in Python and a no-code web application. It comprises tools for data preprocessing, nonlinear feature extraction, machine learning model training, and feature downselection. Users can easily train different types of simple, interpretable ML models on a set of feature transformations quickly and with modest computational resources. In this work, we outline the methods of Spectra-Scope and its effectiveness across diverse datasets, with applications to materials and agricultural spectroscopy data. We show that Spectra-Scope can reproduce performance of comparable models in the literature, and highlight how our emphasis on interpretability can be used to rationalize the behavior of individual models and understand the physical processes behind spectral features. Introduction Spectroscopy is a powerful method for scientific analysis that probes the interaction of matter with electromagnetic radiation. It allows for the investigation of electronic, structural, and dynamic properties of physical systems, and is widely employed across disciplines such as materials science, chemistry, physics, and the life sciences. Autonomous and high throughput laboratories are increasingly popular for exploring the synthesis process space for materials [ 1 , 2 , 3 , 4 , 5 ] . These high-throughput experimental schemes necessitate companion analysis pipelines for characterization of the complex datasets they produce. Automatic spectroscopic analysis tools have thus become a key component of the high-throughput experimentation pipeline. This analysis can span from simple peak fitting [ 6 ] to advanced machine learning (ML) [ 7 , 8 , 9 ] depending on the experimental technique and ultimate goal. Spectroscopic models can enable real-time feedback that supports decision-making during experiments [ 10 ] , enhances the detection of subtle features and patterns in data [ 11 ] , and alleviates the burden of repetitive manual analysis [ 12 ] . Thus, designing robust, interpretable, and computationally efficient models for spectroscopic data analysis is a ubiquitous and important challenge in developing high-throughput experimentation workflows. Automated machine learning (AutoML) software pipelines streamline and automate common ML tasks like data preparation, feature engineering, model training, and hyperparameter tuning. While AutoML is a well-established technique in machine learning communities, few tools exist to apply and optimize such pipelines for materials-specific data. Previous works have applied AutoML techniques to specific materials systems or experimental pipelines, but lack generalizability to other experimental techniques or materials datasets [ 13 , 14 ] . More general purpose frameworks have been established for manufacturing or time-series data [ 15 , 16 , 17 ] . To the best of our knowledge, the literature lacks a versatile toolkit that seamlessly integrates existing methods for automated feature generation, model training, feature downselection, and inference for spectroscopy data. Exposing these functionalities in a common package and interface might help accelerate the model development and data interrogation process for both experts and novice users. This work describes Spectra-Scope, an open-source AutoML framework with a Python and web app implementation that automates simple supervised learning model development from spectral data. It comprises three main capabilities: (i) A library of spectral featurizers that can help transform raw spectra into features that better correlate with target properties, (ii) Model training using random forests [ 18 ] and regularized linear regression, both of which offer handles for interpretability, and (iii) Demonstrated support for different simultaneous modalities of data. The capabilities demonstrated here all only require modest computational resources. The overall workflow is described in Figure 1 . Full details are provided in the Methods section. Figure 1: Outline of this paper and the Spectra-Scope pipeline. (a) Input data can come from any experimental or simulated 1-D array data source for inference on a scalar response variable. (b) Available featurizations of spectral data include the cumulative distribution function, gaussian peak fitting, principal component analysis, polynomial peak fitting, and others as outlined in the methods. (c) Transformed spectra are used to train a machine learning algorithm. This paper focuses on the LCEN algorithm and random forests, but Spectra-Scope can be incorporated with user built models as well. (d) Model training of LCEN and random forests includes feature selection either in the form of LCEN feature downselection or random forest feature importances. (e) The algorithms are used to predict the input response variables. Feature selection helps with model interpretability and investigating modality importance. Methods In this section, we will walk through three main features of the work: (i) the featurization methods we use, (ii) the models that we implement, and (iii) an overview of the application. Featurization We present a bundled implementation of "featurizers" to transform input spectra, extracting features that may better correlate with a property of interest than the raw spectra. This compiles techniques seen across the literature in one place. This allows practitioners to screen over featurizations with no a priori knowledge of the expected relationships between spectral signal and response variables. Examples include transformations such as the cumulative distribution function [ 19 ] and multi-scale polynomial fitting [ 20 ] . We categorize our transformations and features into three conceptual sets: 1. “Local” features (that capture information in finite neighborhoods along the spectrum e.g. polynomial fitting [ 20 ] ), 2. “Nonlocal” features (which incorporate information from the entire spectrum e.g. Fourier transformations), and 3. “Setwise” features, which incorporate information from all spectra in the dataset (e.g. PCA). Local features are useful when spectral features are localized to specific energy regions. They allow for easy interpretation between transformed features and energy regions, which can help identify why specific features may be important for prediction. However, they can miss global information on e.g. periodic or oscillatory trends that nonlocal features can highlight. For this reason we provide both local and nonlocal featurization techniques to complement each other for prediction. Setwise features like PCA create greater risks for extrapolation as they rely on properties of the original training dataset, but may help with understanding global properties of the dataset. We summarize our featurization methods in the Table 1 , and leave full descriptions of them to the supplemental information. Featurization Type Interpolation - Multiscale Polynomial Fitting Local Gaussian Peak Fitting Local Continuous Wavelet Transform Local Nonlinear expansion Local Fourier Transform Nonlocal Cumulative Distribution Function Nonlocal Principal Component Analysis Setwise Feature Agglomeration Setwise Table 1: A brief overview of available featurizations in Spectra-Scope. Explanations of the nomenclature provided in the text. Nonlinear Feature Expansion Because raw spectral values and featurized values may not always have a linear correlation with the response variable, nonlinear feature expansion [ 16 ] can help augment the input variables with a set of nonlinear transformations. For example, x^{2},\ln{x},(\ln{x})^{2},\sqrt{x},\frac{1}{\sqrt{x}},\frac{1}{x},\frac{1}{x^{2}},x^{1.5},\frac{\ln{x}}{x} . Any feature, such

Read the complete text