Generating Flavor Molecules Using Scientific Machine Learning.

Flavor is an essential component in the development of numerous products in the market. The increasing consumption of processed and fast food and healthy packaged food has upraised the investment in new flavoring agents and consequently in molecules with flavoring properties. In this context, this work brings up a scientific machine learning (SciML) approach to address this product engineering need. SciML in computational chemistry has opened paths in the compound's property prediction without requiring synthesis. This work proposes a novel framework of deep generative models within this context to design new flavor molecules. Through the analysis and study of the molecules obtained from the generative model training, it was possible to conclude that even though the generative model designs the molecules through random sampling of actions, it can find molecules that are already used in the food industry, not necessarily as a flavoring agent, or in other industrial sectors. Hence, this corroborates the potential of the proposed methodology for the prospecting of molecules to be applied in the flavor industry.

A Reinforcement Learning Framework to Discover Natural Flavor Molecules.

Flavor is the focal point in the flavor industry, which follows social tendencies and behaviors. The research and development of new flavoring agents and molecules are essential in this field. However, the development of natural flavors plays a critical role in modern society. Considering this, the present work proposes a novel framework based on scientific machine learning to undertake an emerging problem in flavor engineering and industry. It proposes a combining system composed of generative and reinforcement learning models. Therefore, this work brings an innovative methodology to design new flavor molecules. The molecules were evaluated regarding synthetic accessibility, the number of atoms, and the likeness to a natural or pseudo-natural product. This work brings as contributions the implementation of a web scraper code to sample a flavors database and the integration of two scientific machine learning techniques in a complex system as a framework. The implementation of the complex system instead of the generative model by itself obtained 10% more molecules within the optimal results. The designed molecules obtained as an output of the reinforcement learning model's generation were assessed regarding their existence or not in the market and whether they are already used in the flavor industry or not. Thus, we corroborated the potentiality of the framework presented for the search of molecules to be used in the development of flavor-based products.

Novel Framework for Simulated Moving Bed Reactor Optimization Based on Deep Neural Network Models and Metaheuristic Optimizers: An Approach with Optimality Guarantee.

Model-based optimization of simulated moving bed reactors (SMBRs) requires efficient solvers and significant computational power. Over the past years, surrogate models have been considered for such computationally demanding optimization problems. In this sense, artificial neural networks-ANNs-have found applications for modeling the simulated moving bed (SMB) unit but not yet been reported for the reactive SMB (SMBR). Despite ANNs' high accuracy, it is essential to assess its capacity to represent the optimization landscape well. However, a consistent method for optimality assessment using surrogate models is still an open issue in the literature. As such, two main contributions can be highlighted: the SMBR optimization based on deep recurrent neural networks (DRNNs) and the characterization of the feasible operation region. This is done by recycling the data points from a metaheuristic technique-optimality assessment. The results demonstrate that the DRNN-based optimization can address such complex optimization while meeting optimality.