Interpretable deep learning for guided microstructure-property explorations in photovoltaics

The microstructure determines the photovoltaic performance of a thin film organic semiconductor film. The relationship between microstructure and performance is usually highly non-linear and expensive to evaluate, thus making microstructure optimization challenging. Here, we show a data-driven approach for mapping the microstructure to photovoltaic performance using deep convolutional neural networks. We characterize this approach in terms of two critical metrics, its generalizability (has it learnt a reasonable map?), and its intepretability (can it produce meaningful microstructure characteristics that influence its prediction?). A surrogate model that exhibits these two features of generalizability and intepretability is particularly useful for subsequent design exploration. We illustrate this by using the surrogate model for both manual exploration (that verifies known domain insight) as well as automated microstructure optimization. We envision such approaches to be widely applicable to a wide variety of microstructure-sensitive design problems.

[1]  Baskar Ganapathysubramanian,et al.  Encoding Invariances in Deep Generative Models , 2019, ArXiv.

[2]  Olga Wodo,et al.  A graph-based formulation for computational characterization of bulk heterojunction morphology , 2012 .

[3]  J. DiCarlo,et al.  Using goal-driven deep learning models to understand sensory cortex , 2016, Nature Neuroscience.

[4]  B. Ganapathysubramanian,et al.  Sensitivity analysis of current generation in organic solar cells—comparing bilayer, sawtooth, and bulk heterojunction morphologies , 2013 .

[5]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[6]  B. Ganapathysubramanian,et al.  Modeling morphology evolution during solvent-based fabrication of organic solar cells , 2011, 1109.3239.

[7]  Yoshua Bengio,et al.  Convolutional networks for images, speech, and time series , 1998 .

[8]  Abhishek Das,et al.  Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization , 2016, 2017 IEEE International Conference on Computer Vision (ICCV).

[9]  Mark Lundstrom,et al.  Can morphology tailoring improve the open circuit voltage of organic solar cells , 2012 .

[10]  Baskar Ganapathysubramanian,et al.  An explainable deep machine vision framework for plant stress phenotyping , 2018, Proceedings of the National Academy of Sciences.

[11]  Hod Lipson,et al.  Understanding Neural Networks Through Deep Visualization , 2015, ArXiv.

[12]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[13]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[14]  Jean Ponce,et al.  Learning mid-level features for recognition , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[15]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.

[16]  Andrew Zisserman,et al.  Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps , 2013, ICLR.

[17]  Baskar Ganapathysubramanian,et al.  Deep Learning for Flow Sculpting: Insights into Efficient Learning using Scientific Simulation Data , 2017, Scientific Reports.

[18]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[19]  Niyazi Serdar Sariciftci,et al.  Organic solar cells: An overview , 2004 .

[20]  Baskar Ganapathysubramanian,et al.  Algorithm 1025: PARyOpt: A Software for Parallel Asynchronous Remote Bayesian Optimization , 2018, ACM Trans. Math. Softw..

[21]  Baskar Ganapathysubramanian,et al.  Process optimization for microstructure-dependent properties in thin film organic electronics , 2018 .

[22]  Olga Wodo,et al.  Automated, high throughput exploration of process–structure–property relationships using the MapReduce paradigm , 2015 .

[23]  A. Rudin,et al.  Prediction of Flory–Huggins interaction parameters from intrinsic viscosities , 1982 .

[24]  Alán Aspuru-Guzik,et al.  Inverse molecular design using machine learning: Generative models for matter engineering , 2018, Science.

[25]  Darko Stefanovic,et al.  Deep Neural Networks Based Recognition of Plant Diseases by Leaf Image Classification , 2016, Comput. Intell. Neurosci..

[26]  Carlos Guestrin,et al.  "Why Should I Trust You?": Explaining the Predictions of Any Classifier , 2016, ArXiv.

[27]  D. Ginger,et al.  New SPM techniques for analyzing OPV materials , 2010 .

[28]  Wojciech Samek,et al.  Methods for interpreting and understanding deep neural networks , 2017, Digit. Signal Process..

[29]  Baskar Ganapathysubramanian,et al.  Physics-aware Deep Generative Models for Creating Synthetic Microstructures , 2018, ArXiv.

[30]  Sebastian Thrun,et al.  Dermatologist-level classification of skin cancer with deep neural networks , 2017, Nature.

[31]  Fumio Okura,et al.  How Convolutional Neural Networks Diagnose Plant Disease , 2019, Plant phenomics.

[32]  Honglak Lee,et al.  Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations , 2009, ICML '09.

[33]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[34]  Baskar Ganapathysubramanian,et al.  Explaining hyperspectral imaging based plant disease identification: 3D CNN and saliency maps , 2018, ArXiv.

[35]  Guido Raos,et al.  Methodological assessment of kinetic Monte Carlo simulations of organic photovoltaic devices: the treatment of electrostatic interactions. , 2010, The Journal of chemical physics.

[36]  Michael I. Jordan,et al.  The Handbook of Brain Theory and Neural Networks , 2002 .

[37]  R. A. Orwoll.  The Polymer-Solvent Interaction Parameter X , 1977 .

[38]  A. Walker,et al.  Dynamical Monte Carlo modelling of organic solar cells: the dependence of internal quantum efficiency on morphology. , 2005, Nano letters.

[39]  Mohammad Mansouri,et al.  An explainable deep-learning algorithm for the detection of acute intracranial haemorrhage from small datasets , 2018, Nature Biomedical Engineering.

[40]  J. Brédas,et al.  Intermixing at the Pentacene‐Fullerene Bilayer Interface: A Molecular Dynamics Study , 2013, Advanced materials.

[41]  Paul Covington,et al.  Deep Neural Networks for YouTube Recommendations , 2016, RecSys.

[42]  E. Favvas,et al.  What is spinodal decomposition , 2008 .

[43]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[44]  Lingyi Meng,et al.  Dynamic Monte Carlo simulation for highly efficient polymer blend photovoltaics. , 2010, The journal of physical chemistry. B.

[45]  B. Ganapathysubramanian,et al.  Microstructure design using graphs , 2018, npj Computational Materials.

[46]  T. M. Dieb,et al.  Machine Learning-Based Experimental Design in Materials Science , 2018 .

[47]  Richard H. Friend,et al.  Direct Measurement of Electric Field‐Assisted Charge Separation in Polymer:Fullerene Photovoltaic Diodes , 2010, Advanced materials.

[48]  Daniel Moses,et al.  Photoinduced Carrier Generation in P3HT/PCBM Bulk Heterojunction Materials , 2008 .

[49]  Tao Wang,et al.  End-to-end text recognition with convolutional neural networks , 2012, Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012).

[50]  P. Flory.  Thermodynamics of High Polymer Solutions , 1941 .

[51]  Olga Wodo,et al.  Entropy-Isomap: Manifold Learning for High-dimensional Dynamic Processes , 2018, 2018 IEEE International Conference on Big Data (Big Data).

[52]  Radu Timofte,et al.  NTIRE 2018 Challenge on Spectral Reconstruction from RGB Images , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[53]  Baskar Ganapathysubramanian,et al.  Computer simulation of heterogeneous polymer photovoltaic devices , 2012 .

[54]  Avanti Shrikumar,et al.  Learning Important Features Through Propagating Activation Differences , 2017, ICML.

[55]  Junichiro Shiomi,et al.  Designing Nanostructures for Phonon Transport via Bayesian Optimization , 2016, 1609.04972.

[56]  Christian Biemann,et al.  What do we need to build explainable AI systems for the medical domain? , 2017, ArXiv.

[57]  Baskar Ganapathysubramanian,et al.  A computational framework to investigate charge transport in heterogeneous organic photovoltaic devices , 2012 .

[58]  Guigang Zhang,et al.  Deep Learning , 2016, Int. J. Semantic Comput..

[59]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[60]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[61]  Marc'Aurelio Ranzato,et al.  Unsupervised Learning of Invariant Feature Hierarchies with Applications to Object Recognition , 2007, 2007 IEEE Conference on Computer Vision and Pattern Recognition.

[62]  Geoffrey E. Hinton,et al.  Speech recognition with deep recurrent neural networks , 2013, 2013 IEEE International Conference on Acoustics, Speech and Signal Processing.

[63]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[64]  P. Cochat,et al.  Et al , 2008, Archives de pediatrie : organe officiel de la Societe francaise de pediatrie.

[65]  Davide Castelvecchi,et al.  Can we open the black box of AI? , 2016, Nature.

[66]  Adarsh Krishnamurthy,et al.  Learning and Visualizing Localized Geometric Features Using 3D-CNN: An Application to Manufacturability Analysis of Drilled Holes , 2017, ArXiv.

[67]  R. Rubinstein.  The Cross-Entropy Method for Combinatorial and Continuous Optimization , 1999 .

[68]  Adarsh Krishnamurthy,et al.  Learning localized features in 3D CAD models for manufacturability analysis of drilled holes , 2018, Comput. Aided Geom. Des..

[69]  Neil C. Greenham,et al.  A microscopic model for the behavior of nanostructured organic photovoltaic devices , 2007 .

[70]  B. Frey,et al.  Predicting the sequence specificities of DNA- and RNA-binding proteins by deep learning , 2015, Nature Biotechnology.

[71]  Xin Zhang,et al.  End to End Learning for Self-Driving Cars , 2016, ArXiv.