Optimization of Formulations Using Robotic Experiments Driven by Machine Learning DoE

Summary Formulated products are complex mixtures of ingredients whose time to market can be difficult to speed due to the lack of general predictable physical models for the desired properties. Here, we report the coupling of a machine learning classification algorithm with the Thompson sampling efficient multiobjective optimization (TSEMO) algorithm for the simultaneous optimization of continuous and discrete outputs. The methodology is successfully applied to the design of a formulated liquid product of commercial interest for which no physical models are available. Experiments are carried out in a semiautomated fashion using robotic platforms triggered by the machine learning algorithms. The procedure allows one to find nine suitable recipes meeting the customer-defined criteria within 15 working days, outperforming human intuition in the target performance of the formulations.

[1]  Jin Zhao,et al.  Development of Surfactants and Builders in Detergent Formulations , 2008 .

[2]  Rafiqul Gani,et al.  Product design - Molecules, devices, functional products, and formulated products , 2015, Comput. Chem. Eng..

[3]  Gang Chen,et al.  MULTI-CLASS SUPPORT VECTOR MACHINE ACTIVE LEARNING FOR MUSIC ANNOTATION , 2010 .

[4]  Mohammad Reza Banan,et al.  A new PSO-based algorithm for multi-objective optimization with continuous and discrete design variables , 2018 .

[5]  Alexandr Zubov,et al.  Closed-Loop Multitarget Optimization for Discovery of New Emulsion Polymerization Recipes , 2015, Organic process research & development.

[6]  Aleksander Bajc,et al.  Use of factorial design for evaluation of factors affecting the chemical stability of sirolimus (rapamycin) in solid dosage form. , 2012, Acta chimica Slovenica.

[7]  S. Zendehboudi,et al.  A Comprehensive Review on Emulsions and Emulsion Stability in Chemical and Energy Industries , 2018, The Canadian Journal of Chemical Engineering.

[8]  Leroy Cronin,et al.  Organic synthesis in a modular robotic system driven by a chemical programming language , 2019, Science.

[9]  Hod Lipson,et al.  Distilling Free-Form Natural Laws from Experimental Data , 2009, Science.

[10]  Neal G. Anderson,et al.  Design of Experiments (DoE) and Process Optimization. A Review of Recent Publications , 2015 .

[11]  G. Ronning.  Maximum likelihood estimation of dirichlet distributions , 1989 .

[12]  Rafiqul Gani,et al.  Design of Formulated Products: A Systematic Methodology , 2011 .

[13]  Edward L Cussler,et al.  Chemical product design , 2001 .

[14]  R. Gani,et al.  A comprehensive framework for surfactant selection and design for emulsion based chemical product design , 2014 .

[15]  E. Hines,et al.  Combining Gaussian processes, mutual information and a genetic algorithm for multi-target optimization of expensive-to-evaluate functions , 2014 .

[16]  Lorenzo Bruzzone,et al.  Active and Semisupervised Learning for the Classification of Remote Sensing Images , 2014, IEEE Transactions on Geoscience and Remote Sensing.

[17]  Raquel Costa,et al.  Product design and engineering — past, present, future trends in teaching, research and practices: academic and industry points of view , 2020 .

[18]  S. Neau,et al.  A Quality by Experimental Design Approach to Assess the Effect of Formulation and Process Variables on the Extrusion and Spheronization of Drug-Loaded Pellets Containing Polyplasdone® XL-10 , 2016, AAPS PharmSciTech.

[19]  Harry Zhang,et al.  The Optimality of Naive Bayes , 2004, FLAIRS.

[20]  S. Derkach.  Rheology of emulsions. , 2009, Advances in colloid and interface science.

[21]  Pritam Ranjan,et al.  Space-filling Latin hypercube designs based on randomization restrictions in factorial experiments , 2013 .

[22]  Brian A. Taylor,et al.  Algorithms for the self-optimisation of chemical reactions , 2019, Reaction Chemistry & Engineering.

[23]  Hao Wang,et al.  A Multicriteria Generalization of Bayesian Global Optimization , 2016, Advances in Stochastic and Deterministic Global Optimization.

[24]  Kalyanmoy Deb,et al.  A fast and elitist multiobjective genetic algorithm: NSGA-II , 2002, IEEE Trans. Evol. Comput..

[25]  C. E. SHANNON,et al.  A mathematical theory of communication , 1948, MOCO.

[26]  Vassilios S. Vassiliadis,et al.  A new formulation for symbolic regression to identify physico-chemical laws from experimental data , 2020, Chemical Engineering Journal.

[27]  Artur M. Schweidtmann,et al.  Machine learning meets continuous flow chemistry: Automated optimization towards the Pareto front of multiple objectives , 2018, Chemical Engineering Journal.

[28]  E. Park,et al.  Quality by Design (QbD) approach to optimize the formulation of a bilayer combination tablet (Telmiduo®) manufactured via high shear wet granulation. , 2017, International journal of pharmaceutics.

[29]  Ravikanth Kona,et al.  Quality by Design I: Application of Failure Mode Effect Analysis (FMEA) and Plackett–Burman Design of Experiments in the Identification of “Main Factors” in the Formulation and Process Design Space for Roller-Compacted Ciprofloxacin Hydrochloride Immediate-Release Tablets , 2012, AAPS PharmSciTech.

[30]  M. Gohel,et al.  Formulation optimization of controlled release diclofenac sodium microspheres using factorial design. , 1998, Journal of controlled release : official journal of the Controlled Release Society.

[31]  Jerome Sacks,et al.  Choosing the Sample Size of a Computer Experiment: A Practical Guide , 2009, Technometrics.

[32]  Fabio Fiorelli,et al.  Deep learning‐based surrogate modeling and optimization for microalgal biofuel production and photobioreactor design , 2018, AIChE Journal.

[33]  Leroy Cronin,et al.  Controlling an organic synthesis robot with machine learning to search for new reactivity , 2018, Nature.

[34]  V. Venkatasubramanian.  The promise of artificial intelligence in chemical engineering: Is it here, finally? , 2018, AIChE Journal.

[35]  Lawrence X. Yu,et al.  Quality by design approach for understanding the critical quality attributes of cyclosporine ophthalmic emulsion. , 2014, Molecular pharmaceutics.

[36]  S. Savić,et al.  An Overview of Novel Surfactants for Formulation of Cosmetics with Certain Emphasis on Acidic Active Substances , 2016 .

[37]  Pedro M. Saraiva,et al.  Chemical product engineering: An emerging paradigm within chemical engineering , 2006 .

[38]  R. Pal.  Viscosity models for multiple emulsions , 2008 .

[39]  Klavs F. Jensen,et al.  Suzuki–Miyaura cross-coupling optimization enabled by automated feedback , 2016, Reaction chemistry & engineering.

[40]  Rafiqul Gani,et al.  A Systematic Methodology for Design of Emulsion Based Chemical Products , 2012 .

[41]  Artur M. Schweidtmann,et al.  Machine learning and molecular descriptors enable rational solvent selection in asymmetric catalysis , 2019, Chemical science.

[42]  Mouloud Koudil,et al.  A Novel Active Learning Method Using SVM for Text Classification , 2018, Int. J. Autom. Comput..

[43]  Wei Chen,et al.  Bayesian Optimization for Materials Design with Mixed Quantitative and Qualitative Variables , 2019, Scientific Reports.

[44]  R. Gani,et al.  Computer aided product design: problem formulations, methodology and applications , 1996 .

[45]  Tadayoshi Fushiki,et al.  Estimation of prediction error by using K-fold cross-validation , 2011, Stat. Comput..

[46]  De-Liang Long,et al.  A Modular Programmable Inorganic Cluster Discovery Robot for the Discovery and Synthesis of Polyoxometalates , 2020, ACS central science.

[47]  Artur M. Schweidtmann,et al.  Efficient multiobjective optimization employing Gaussian processes, spectral sampling and a genetic algorithm , 2018, Journal of Global Optimization.

[48]  S. S. Imam,et al.  Nanostructured lipid carriers of pioglitazone for transdermal application: from experimental design to bioactivity detail , 2016, Drug delivery.

[49]  Klavs F. Jensen,et al.  Photoredox Iridium–Nickel Dual-Catalyzed Decarboxylative Arylation Cross-Coupling: From Batch to Continuous Flow via Self-Optimizing Segmented Flow Reactor , 2018 .

[50]  William A. Gale,et al.  A sequential algorithm for training text classifiers , 1994, SIGIR '94.

[51]  Joshua D. Knowles,et al.  ParEGO: a hybrid algorithm with on-line landscape approximation for expensive multiobjective optimization problems , 2006, IEEE Transactions on Evolutionary Computation.

[52]  R. Marler,et al.  The weighted sum method for multi-objective optimization: new insights , 2010 .

[53]  Max Tegmark,et al.  AI Feynman: A physics-inspired method for symbolic regression , 2019, Science Advances.

[54]  Claudio Battilocchio,et al.  A Novel Internet-Based Reaction Monitoring, Control and Autonomous Self-Optimization Platform for Chemical Synthesis , 2015 .

[55]  H. Schubert,et al.  Product and Formulation Engineering of Emulsions , 2004 .

[56]  Cláudia Silva,et al.  Quality by Design in Pharmaceutical Manufacturing: a systematic review of current status, challenges and future perspectives. , 2019, European journal of pharmaceutics and biopharmaceutics : official journal of Arbeitsgemeinschaft fur Pharmazeutische Verfahrenstechnik e.V.