arXiv subjects

Savitha Ramasamy

Publications and source records attributed to Savitha Ramasamy.

23 records · Page 2Linked to original sources

arXiv2019

A novel method for extracting interpretable knowledge from a spiking neural classifier with time-varying synaptic weights

This paper presents a novel method for information interpretability in an MC-SEFRON classifier. To develop a method to extract knowledge stored in a trained classifier, first, the binary-class SEFRON classifier developed earlier is extended to handle multi-class problems. MC-SEFRON uses the population encoding scheme to encode the real-valued input data into spike patterns. MC-SEFRON is trained using the same supervised learning rule used in the SEFRON. After training, the proposed method extracts the knowledge for a given class stored in the classifier by mapping the weighted postsynaptic potential in the time domain to the feature domain as Feature Strength Functions (FSFs). A set of FSFs corresponding to each output class represents the extracted knowledge from the classifier. This knowledge encoding method is derived to maintain consistency between the classification in the time domain and the feature domain. The correctness of the FSF is quantitatively measured by using FSF directly for classification tasks. For a given input, each FSF is sampled at the input value to obtain the corresponding feature strength value (FSV). Then the aggregated FSVs obtained for each class are used to determine the output class labels during classification. FSVs are also used to interpret the predictions during the classification task. Using ten UCI datasets and the MNIST dataset, the knowledge extraction method, interpretation and the reliability of the FSF are demonstrated. Based on the studies, it can be seen that on an average, the difference in the classification accuracies using the FSF directly and those obtained by MC-SEFRON is only around 0.9% & 0.1\% for the UCI datasets and the MNIST dataset respectively. This clearly shows that the knowledge represented by the FSFs has acceptable reliability and the interpretability of classification using the classifier's knowledge has been justified.

cs.NE↗

arXiv2018

Accelerating Photovoltaic Materials Development via High-Throughput Experiments and Machine-Learning-Assisted Diagnosis

Accelerating the experimental cycle for new materials development is vital for addressing the grand energy challenges of the 21st century. We fabricate and characterize 75 unique halide perovskite-inspired solution-based thin-film materials within a two-month period, with 87% exhibiting band gaps between 1.2 eV and 2.4 eV that are of interest for energy-harvesting applications. This increased throughput is enabled by streamlining experimental workflows, developing a set of precursors amenable to high-throughput synthesis, and developing machine-learning assisted diagnosis. We utilize a deep neural network to classify compounds based on experimental X-ray diffraction data into 0D, 2D, and 3D structures more than 10 times faster than human analysis and with 90% accuracy. We validate our methods using lead-halide perovskites and extend the application to novel lead-free compositions. The wider synthesis window and faster cycle of learning enables three noteworthy scientific findings: (1) we realize four inorganic layered perovskites, A3B2Br9 (A = Cs, Rb; B = Bi, Sb) in thin-film form via one-step liquid deposition; (2) we report a multi-site lead-free alloy series that was not previously described in literature, Cs3(Bi1-xSbx)2(I1-xBrx)9; and (3) we reveal the effect on bandgap (reduction to <2 eV) and structure upon simultaneous alloying on the B-site and X-site of Cs3Bi2I9 with Sb and Br. This study demonstrates that combining an accelerated experimental cycle of learning and machine-learning based diagnosis represents an important step toward realizing fully-automated laboratories for materials discovery and development.

physics.app-ph↗

arXiv2018

Predicting thermoelectric properties from crystal graphs and material descriptors - first application for functional materials

We introduce the use of Crystal Graph Convolutional Neural Networks (CGCNN), Fully Connected Neural Networks (FCNN) and XGBoost to predict thermoelectric properties. The dataset for the CGCNN is independent of Density Functional Theory (DFT) and only relies on the crystal and atomic information, while that for the FCNN is based on a rich attribute list mined from Materialsproject.org. The results show that the optimized FCNN is three layer deep and is able to predict the scattering-time independent thermoelectric powerfactor much better than the CGCNN (or XGBoost), suggesting that bonding and density of states descriptors informed from materials science knowledge obtained partially from DFT are vital to predict functional properties.

physics.comp-ph↗

arXiv2018

Autonomous Deep Learning: Incremental Learning of Denoising Autoencoder for Evolving Data Streams

The generative learning phase of Autoencoder (AE) and its successor Denosing Autoencoder (DAE) enhances the flexibility of data stream method in exploiting unlabelled samples. Nonetheless, the feasibility of DAE for data stream analytic deserves in-depth study because it characterizes a fixed network capacity which cannot adapt to rapidly changing environments. An automated construction of a denoising autoeconder, namely deep evolving denoising autoencoder (DEVDAN), is proposed in this paper. DEVDAN features an open structure both in the generative phase and in the discriminative phase where input features can be automatically added and discarded on the fly. A network significance (NS) method is formulated in this paper and is derived from the bias-variance concept. This method is capable of estimating the statistical contribution of the network structure and its hidden units which precursors an ideal state to add or prune input features. Furthermore, DEVDAN is free of the problem- specific threshold and works fully in the single-pass learning fashion. The efficacy of DEVDAN is numerically validated using nine non-stationary data stream problems simulated under the prequential test-then-train protocol where DEVDAN is capable of delivering an improvement of classification accuracy to recently published online learning works while having flexibility in the automatic extraction of robust input features and in adapting to rapidly changing environments.

cs.LG↗

arXiv2018

Online Deep Learning: Growing RBM on the fly

We propose a novel online learning algorithm for Restricted Boltzmann Machines (RBM), namely, the Online Generative Discriminative Restricted Boltzmann Machine (OGD-RBM), that provides the ability to build and adapt the network architecture of RBM according to the statistics of streaming data. The OGD-RBM is trained in two phases: (1) an online generative phase for unsupervised feature representation at the hidden layer and (2) a discriminative phase for classification. The online generative training begins with zero neurons in the hidden layer, adds and updates the neurons to adapt to statistics of streaming data in a single pass unsupervised manner, resulting in a feature representation best suited to the data. The discriminative phase is based on stochastic gradient descent and associates the represented features to the class labels. We demonstrate the OGD-RBM on a set of multi-category and binary classification problems for data sets having varying degrees of class-imbalance. We first apply the OGD-RBM algorithm on the multi-class MNIST dataset to characterize the network evolution. We demonstrate that the online generative phase converges to a stable, concise network architecture, wherein individual neurons are inherently discriminative to the class labels despite unsupervised training. We then benchmark OGD-RBM performance to other machine learning, neural network and ClassRBM techniques for credit scoring applications using 3 public non-stationary two-class credit datasets with varying degrees of class-imbalance. We report that OGD-RBM improves accuracy by 2.5-3% over batch learning techniques while requiring at least 24%-70% fewer neurons and fewer training samples. This online generative training approach can be extended greedily to multiple layers for training Deep Belief Networks in non-stationary data mining applications without the need for a priori fixed architectures.

cs.NE↗