Deep learning using multilayer perception improves the diagnostic acumen of spirometry: a single-centre Canadian study.

RATIONALE: Spirometry and plethysmography are the gold standard pulmonary function tests (PFT) for diagnosis and management of lung disease. Due to the inaccessibility of plethysmography, spirometry is often used alone but this leads to missed or misdiagnoses as spirometry cannot identify restrictive disease without plethysmography. We aimed to develop a deep learning model to improve interpretation of spirometry alone. METHODS: We built a multilayer perceptron model using full PFTs from 748 patients, interpreted according to international guidelines. Inputs included spirometry (forced vital capacity, forced expiratory volume in 1 s, forced mid-expiratory flow25-75), plethysmography (total lung capacity, residual volume) and biometrics (sex, age, height). The model was developed with 2582 PFTs from 477 patients, randomly divided into training (80%), validation (10%) and test (10%) sets, and refined using 1245 previously unseen PFTs from 271 patients, split 50/50 as validation (136 patients) and test (135 patients) sets. Only one test per patient was used for each of 10 experiments conducted for each input combination. The final model was compared with interpretation of 82 spirometry tests by 6 trained pulmonologists and a decision tree. RESULTS: Accuracies from the first 477 patients were similar when inputs included biometrics+spirometry+plethysmography (95%±3%) vs biometrics+spirometry (90%±2%). Model refinement with the next 271 patients improved accuracies with biometrics+pirometry (95%±2%) but no change for biometrics+spirometry+plethysmography (95%±2%). The final model significantly outperformed (94.67%±2.63%, p<0.01 for both) interpretation of 82 spirometry tests by the decision tree (75.61%±0.00%) and pulmonologists (66.67%±14.63%). CONCLUSIONS: Deep learning improves the diagnostic acumen of spirometry and classifies lung physiology better than pulmonologists with accuracies comparable to full PFTs.

EDropout: Energy-Based Dropout and Pruning of Deep Neural Networks.

Dropout is a well-known regularization method by sampling a sub-network from a larger deep neural network and training different sub-networks on different subsets of the data. Inspired by the dropout concept, we propose EDropout as an energy-based framework for pruning neural networks in classification tasks. In this approach, a set of binary pruning state vectors (population) represents a set of corresponding sub-networks from an arbitrary original neural network. An energy loss function assigns a scalar energy loss value to each pruning state. The energy-based model (EBM) stochastically evolves the population to find states with lower energy loss. The best pruning state is then selected and applied to the original network. Similar to dropout, the kept weights are updated using backpropagation in a probabilistic model. The EBM again searches for better pruning states and the cycle continuous. This procedure is a switching between the energy model, which manages the pruning states, and the probabilistic model, which updates the kept weights, in each iteration. The population can dynamically converge to a pruning state. This can be interpreted as dropout leading to pruning the network. From an implementation perspective, unlike most of the pruning methods, EDropout can prune neural networks without manually modifying the network architecture code. We have evaluated the proposed method on different flavors of ResNets, AlexNet, l1 pruning, ThinNet, ChannelNet, and SqueezeNet on the Kuzushiji, Fashion, CIFAR-10, CIFAR-100, Flowers, and ImageNet data sets, and compared the pruning rate and classification performance of the models. The networks trained with EDropout on average achieved a pruning rate of more than 50% of the trainable parameters with approximately < 5% and < 1% drop of Top-1 and Top-5 classification accuracy, respectively.

Robust Distributed Collaborative Beamforming for Wireless Sensor Networks with Channel Estimation Impairments.

We propose a new collaborative beamforming (CB) solution robust (i.e., RCB) against major channel estimation impairments over dual-hop transmissions through a wireless sensor network (WSN) of K nodes. The source first sends its signal to the WSN. Then, each node forwards its received signal after multiplying it by a properly selected beamforming weight. The latter aims to minimize the received noise power while maintaining the desired power equal to unity. These weights depend on some channel state information (CSI) parameters. Hence, they have to be estimated locally at each node, thereby, resulting in channel estimation errors that could severely hinder CB performance. Exploiting an efficient asymptotic approximation at large K, we develop alternative RCB solutions that adapt to different implementation scenarios and wireless propagation environments ranging from monochromatic (i.e., scattering-free) to polychromatic (i.e., scattered) ones. Besides, in contrast to existing techniques, our new RCB solutions are distributed (i.e., DCB) in that they do not require any information exchange among nodes, thereby dramatically improving both WSN spectral and power efficiencies. Simulation results confirm that the proposed robust DCB (RDCB) techniques are much more robust in terms of achieved signal-to-noise ratio (SNR) against channel estimation errors than best representative CB benchmarks.

Synthesizing Chest X-Ray Pathology for Training Deep Convolutional Neural Networks.

Medical datasets are often highly imbalanced with over-representation of prevalent conditions and poor representation of rare medical conditions. Due to privacy concerns, it is challenging to aggregate large datasets between health care institutions. We propose synthesizing pathology in medical images as a means to overcome these challenges. We implement a deep convolutional generative adversarial network (DCGAN) to create synthesized chest X-rays based upon a modest sized labeled dataset. We used a combination of real and synthesized images to train deep convolutional neural networks (DCNNs) to detect pathology across five classes of chest X-rays. The comparative study of DCNNs trained with the combination of real and synthesized images showed that these networks can outperform similar networks trained solely with real images in pathology classification. This improved performance is largely attributable to the balancing of the dataset using DCGAN synthesized images, where classes that are lacking in example images are preferentially augmented.

On the relationship between multicast/broadcast throughput and resource utilizations in wireless mesh networks.

This paper deals with the problem of multicast/broadcast throughput in multi-channel multi-radio wireless mesh networks that suffer from the resource constraints. We provide a formulation to capture the utilization of the network resources and derive analytical relationships for the network's throughput in terms of the node utilization, the channel utilization, and the number of transmissions. Our model relies on the on-demand quality of service multicast/broadcast sessions, where each admitted session creates a unique tree with a specific bandwidth. As an advantage, the derived relationships are independent of the type of tree built for each session and can be used for different protocols. The proposed formulation considers the channel assignment strategy and reflects both the wireless broadcast advantage and the interference constraint. We also offer a comprehensive discussion to evaluate the effects of load-balancing and number of transmissions on the network's throughput. Numerical results confirm the accuracy of the presented analysis.