Neural network-derived perfusion maps: A model-free approach to computed tomography perfusion in patients with acute ischemic stroke.

Objective: In this study, we investigate whether a Convolutional Neural Network (CNN) can generate informative parametric maps from the pre-processed CT perfusion data in patients with acute ischemic stroke in a clinical setting. Methods: The CNN training was performed on a subset of 100 pre-processed perfusion CT dataset, while 15 samples were kept for testing. All the data used for the training/testing of the network and for generating ground truth (GT) maps, using a state-of-the-art deconvolution algorithm, were previously pre-processed using a pipeline for motion correction and filtering. Threefold cross validation had been used to estimate the performance of the model on unseen data, reporting Mean Squared Error (MSE). Maps accuracy had been checked through manual segmentation of infarct core and total hypo-perfused regions on both CNN-derived and GT maps. Concordance among segmented lesions was assessed using the Dice Similarity Coefficient (DSC). Correlation and agreement among different perfusion analysis methods were evaluated using mean absolute volume differences, Pearson correlation coefficients, Bland-Altman analysis, and coefficient of repeatability across lesion volumes. Results: The MSE was very low for two out of three maps, and low in the remaining map, showing good generalizability. Mean Dice scores from two different raters and the GT maps ranged from 0.80 to 0.87. Inter-rater concordance was high, and a strong correlation was found between lesion volumes of CNN maps and GT maps (0.99, 0.98, respectively). Conclusion: The agreement between our CNN-based perfusion maps and the state-of-the-art deconvolution-algorithm perfusion analysis maps, highlights the potential of machine learning methods applied to perfusion analysis. CNN approaches can reduce the volume of data required by deconvolution algorithms to estimate the ischemic core, and thus might allow the development of novel perfusion protocols with lower radiation dose deployed to the patient.

SeReNe: Sensitivity-Based Regularization of Neurons for Structured Sparsity in Neural Networks.

Deep neural networks include millions of learnable parameters, making their deployment over resource-constrained devices problematic. Sensitivity-based regularization of neurons (SeReNe) is a method for learning sparse topologies with a structure, exploiting neural sensitivity as a regularizer. We define the sensitivity of a neuron as the variation of the network output with respect to the variation of the activity of the neuron. The lower the sensitivity of a neuron, the less the network output is perturbed if the neuron output changes. By including the neuron sensitivity in the cost function as a regularization term, we are able to prune neurons with low sensitivity. As entire neurons are pruned rather than single parameters, practical network footprint reduction becomes possible. Our experimental results on multiple network architectures and datasets yield competitive compression ratios with respect to state-of-the-art references.

LOss-Based SensiTivity rEgulaRization: Towards deep sparse neural networks.

LOBSTER (LOss-Based SensiTivity rEgulaRization) is a method for training neural networks having a sparse topology. Let the sensitivity of a network parameter be the variation of the loss function with respect to the variation of the parameter. Parameters with low sensitivity, i.e. having little impact on the loss when perturbed, are shrunk and then pruned to sparsify the network. Our method allows to train a network from scratch, i.e. without preliminary learning or rewinding. Experiments on multiple architectures and datasets show competitive compression ratios with minimal computational overhead.

Unveiling COVID-19 from CHEST X-Ray with Deep Learning: A Hurdles Race with Small Data.

The possibility to use widespread and simple chest X-ray (CXR) imaging for early screening of COVID-19 patients is attracting much interest from both the clinical and the AI community. In this study we provide insights and also raise warnings on what is reasonable to expect by applying deep learning to COVID classification of CXR images. We provide a methodological guide and critical reading of an extensive set of statistical results that can be obtained using currently available datasets. In particular, we take the challenge posed by current small size COVID data and show how significant can be the bias introduced by transfer-learning using larger public non-COVID CXR datasets. We also contribute by providing results on a medium size COVID CXR dataset, just collected by one of the major emergency hospitals in Northern Italy during the peak of the COVID pandemic. These novel data allow us to contribute to validate the generalization capacity of preliminary results circulating in the scientific community. Our conclusions shed some light into the possibility to effectively discriminate COVID using CXR.

Role of Synaptic Stochasticity in Training Low-Precision Neural Networks.

Stochasticity and limited precision of synaptic weights in neural network models are key aspects of both biological and hardware modeling of learning processes. Here we show that a neural network model with stochastic binary weights naturally gives prominence to exponentially rare dense regions of solutions with a number of desirable properties such as robustness and good generalization performance, while typical solutions are isolated and hard to find. Binary solutions of the standard perceptron problem are obtained from a simple gradient descent procedure on a set of real values parametrizing a probability distribution over the binary synapses. Both analytical and numerical results are presented. An algorithmic extension that allows to train discrete deep neural networks is also investigated.