Here we present Multiagent consensus equilibrium (MACE) can also be demonstrated for both molecular structure determination and spectroscopy dimension reduction.<br/>Spectroscopy dimension reduction analysis is often used for classification of chemicals. Here we implemented MACE to combine different dimension reduction method (e.g. PCA, PCALDA and ensemble method), and found out MACE can systematically merge multiple dimension reduction methods for spectra analysis. We also addressed the overfitting problem of ill-posed LDA, with development generative adversarial linear discriminant analysis (GALDA) .  MACE framework can combine GALDA and PCA to adjust the separation of each class of data (defined as resolution) and overfitting (defined as separation performance difference between training and testing), so that the overfitting problem in dimension reduction should be reduced to a small value, while the resolution is retailed at high value. Initial demonstration of GALDA MACE method for Raman Spectra dimension reduction and classification is presented, along with a critical discussion of the figures of merit and future developments.<br/>In details of molecular structure determination, we discovered MACE can be used in integrating experimental observables as constraints in molecular structure determination and for the systematic merging of multiple computational architectures. MACE can also merge experimental measurement and chemical modeling. Fusing a model of moment of inertia with low-cost computational methods yielded structure determination performance comparable to standard computational.

Multiagent Consensus Equilibrium (MACE) in molecular structure determination and molecular spectral analysis

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.