Predicting the maximum absorption wavelength of azo dyes using an interpretable machine learning strategy

Feature Selection in Machine Learning for Perovskite Materials Design and Discovery

“…The three major scenarios of model application are high-throughput screening (HTS), inverse design, and the development of online prediction programs. HTS uses the constructed model to predict the target variables of a huge number of virtual samples in order to filter out samples with high performance potential and guide experimental synthesis [29,30]. The inverse design can be used to obtain the features of designed samples via the inverse projection method, which is an effective way to realize the material from properties to composition [31,32].…”

Artificial Intelligence for Surface‐Enhanced Raman Spectroscopy

“…XGBoost [97] IR and Raman spectra Predict surface-adsorbate interaction properties ET, [98] SISSO [98] Synthesis Selected synthesis descriptors Classify selected features of spectra ET [99] The citrate to gold (III) ratio, scanning velocity and radiation intensity Predict nanoparticle size ANN [ 100] Spectroscopic characteristics based on UV-vis/DLS Optimize experimental/reaction conditions GA, [ 101] BO+DNN [ 102] Reaction conditions Predict selected spectroscopic characteristics of nanoparticles SVR [103] Target molecules Propose a sequence of chemically viable reaction steps ANN [ 104] Instrumentations and spectral preprocessing Complex 2D images Optimize illumination light source parameters CNN [ 105] Patterns of weakly scattering perturbations Design transmission matrices VAE [ 106] Scattering spectra of nanostructures Encode up to 9 bits of information for high-density optical information storage CNN [ 107] Noisy Raman spectra Remove the baseline, cosmic rays, and noise simultaneously or separately CNN, [108][109][110] ResNet, [ 111] U-net, [ 111] ANN+U-net, [ 112] PCA, [113] GAN [ 114] Spectral analysis SERS spectrum Identify the existence of the molecular fingerprints…”

“…and Mai et al applied XGBoost to predict the maximum absorption wavelength of azo dyes and incorporated SHAP to explain the relationship between the molecular structure and the absorption, [97] both exhibiting the potential in screening suitable molecules upon certain incident wavelength and/or substrate plasmonic on-resonance wavelength. When taking the whole SERS nanotag into consideration, Wang et al leveraged AI to further solve the interaction properties of substrate-adsorbate systems, including adsorption energy, charge transfer, and molecular bond energy, [98] exerting high potential in reducing redundant manual trials and designing high-performance nanotags with reporter molecules meeting criteria of high absorption, large cross section, and efficient charge transfer.…”

“…Specifically, the resonance of the reporter molecules, the charge transfer between the reporters and the SERS substrates, as well as the on‐resonance EM enhancement of the SERS substrates can be coupled within one SERS nanotag to give rise to ultra‐strong signals (i.e., λ reporter ≈ λ substrate ≈ λ charge transfer ≈ λ incidence ). [ 210 ] Though it is hard for conventional wisdoms to tackle such intertwining problems, in terms of the fluorescence molecules only, Joung et al used a DL model to solve the properties of fluorescence including absorption and emission peak wavelengths in different solvents, [ 188 ] and Mai et al applied XGBoost to predict the maximum absorption wavelength of azo dyes and incorporated SHAP to explain the relationship between the molecular structure and the absorption, [ 97 ] both exhibiting the potential in screening suitable molecules upon certain incident wavelength and/or substrate plasmonic on‐resonance wavelength. When taking the whole SERS nanotag into consideration, Wang et al leveraged AI to further solve the interaction properties of substrate‐adsorbate systems, including adsorption energy, charge transfer, and molecular bond energy, [ 98 ] exerting high potential in reducing redundant manual trials and designing high‐performance nanotags with reporter molecules meeting criteria of high absorption, large cross section, and efficient charge transfer.…”

Enhanced Predictions for the Experimental Photophysical Data Using the Featurized Schnet-Bondstep Approach

“…However, the complexity of excited state PESs and the relevant computational cost often hinder the theoretical explorations on these novel molecular properties. Consequently, developing the efficient and reliable computational tools to assess the experimental observables became an increasing interest for the applications like molecular chromophores, − organic solar cell, ,,− photoswitch, and optical materials , by the computational chemistry community.…”