This study compares the relative utility of deep learning models as automated phenotypic classifiers, built with features of peripheral blood cell populations assayed with flow cytometry. We report surprisingly good sensitively for the task of distinguishing male from female mice solely from attributes of cells present in peripheral blood samples. This novel application of cytometry technology is easily validated and represents an example use that is not reliably possible with current visual analysis techniques. It serves as a useful test case with known ground truth labels of a subtle cellular phenotype and indicates the potential of this approach to more biologically meaningful phenotypes. We evaluate three different deep neural network architectures that vary in how the high-dimensional flow cytometry data is processed as input: 1) a convolutional neural network applied to the raw, per-cell marker intensity values, 2) a multi-layer perceptron taking per-channel histograms representing the range of intensities for a given marker, and 3) a multi-layer perceptron using a hypervoxel representation of the set of marker intensities for each cell in a given input sample. Our results show the histogram-based multi-layer perceptron achieves the highest classification accuracy of 91% on 2300 mouse blood samples for a mutant phenotype classification task. This indicates the value of deep learning models for detection of subtle signals that are not readily observed in high dimensions by visual analysis. These results demonstrate the effectiveness of deep learning for automated analysis of flow cytometry data on a challenging and biologically complex classification task.

Deep learning is a powerful technique for the analysis of remote sensing imagery. For applications that require real-time processing on mobile platforms, a low power consumption processing unit is advantageous. The human brain is remarkably powerful at image recognition tasks while operating at very low power consumption levels. Neuromorphic computing designs aim to achieve energy efficiency through the use of spiking neurons and low-precision synapses to perform data processing. We demonstrate here the classification of red, green, blue and depth and hyperspectral data sets using a neuromorphic processing unit (IBM TrueNorth Neurosynaptic System). The convolutional neural-network architecture of the classifier network has been adapted to fit the neuromorphic architecture. The results on overhead imagery and hyperspectral imagery data show that neuromorphic platforms can achieve the state-of-theart performance in semantic labeling with significantly (≈1000×) lower power consumption than traditional GPU-based solutions.

Deep Neural Networks (DNN) achieve human level performance in many image analytics tasks but DNNs are mostly deployed to GPU platforms that consume a considerable amount of power. New hardware platforms using lower precision arithmetic achieve drastic reductions in power consumption. More recently, brain-inspired spiking neuromorphic chips have achieved even lower power consumption, on the order of milliwatts, while still offering real-time processing. However, for deploying DNNs to energy efficient neuromorphic chips the incompatibility between continuous neurons and synaptic weights of traditional DNNs, discrete spiking neurons and synapses of neuromorphic chips need to be overcome. Previous work has achieved this by training a network to learn continuous probabilities, before it is deployed to a neuromorphic architecture, such as IBM TrueNorth Neurosynaptic System, by random sampling these probabilities. The main contribution of this paper is a new learning algorithm that learns a TrueNorth configuration ready for deployment. We achieve this by training directly a binary hardware crossbar that accommodates the TrueNorth axon configuration constrains and we propose a different neuron model. Results of our approach trained on electroencephalogram (EEG) data show a significant improvement with previous work (76% vs 86% accuracy) while maintaining state of the art performance on the MNIST handwritten data set.