Real-time data analysis for medical diagnosis using FPGA-accelerated neural networks.

BACKGROUND: Real-time analysis of patient data during medical procedures can provide vital diagnostic feedback that significantly improves chances of success. With sensors becoming increasingly fast, frameworks such as Deep Neural Networks are required to perform calculations within the strict timing constraints for real-time operation. However, traditional computing platforms responsible for running these algorithms incur a large overhead due to communication protocols, memory accesses, and static (often generic) architectures. In this work, we implement a low-latency Multi-Layer Perceptron (MLP) processor using Field Programmable Gate Arrays (FPGAs). Unlike CPUs and Graphics Processing Units (GPUs), our FPGA-based design can directly interface sensors, storage devices, display devices and even actuators, thus reducing the delays of data movement between ports and compute pipelines. Moreover, the compute pipelines themselves are tailored specifically to the application, improving resource utilization and reducing idle cycles. We demonstrate the effectiveness of our approach using mass-spectrometry data sets for real-time cancer detection. RESULTS: We demonstrate that correct parameter sizing, based on the application, can reduce latency by 20% on average. Furthermore, we show that in an application with tightly coupled data-path and latency constraints, having a large amount of computing resources can actually reduce performance. Using mass-spectrometry benchmarks, we show that our proposed FPGA design outperforms both CPU and GPU implementations, with an average speedup of 144x and 21x, respectively. CONCLUSION: In our work, we demonstrate the importance of application-specific optimizations in order to minimize latency and maximize resource utilization for MLP inference. By directly interfacing and processing sensor data with ultra-low latency, FPGAs can perform real-time analysis during procedures and provide diagnostic feedback that can be critical to achieving higher percentages of successful patient outcomes.

Decoupled graph knowledge distillation: A general logits-based method for learning MLPs on graphs.

While Graph Neural Networks (GNNs) have demonstrated their effectiveness in processing non-Euclidean structured data, the neighborhood fetching of GNNs is time-consuming and computationally intensive, making them difficult to deploy in low-latency industrial applications. To address the issue, a feasible solution is graph knowledge distillation (KD), which can learn high-performance student Multi-layer Perceptrons (MLPs) to replace GNNs by mimicking the superior output of teacher GNNs. However, state-of-the-art graph knowledge distillation methods are mainly based on distilling deep features from intermediate hidden layers, this leads to the significance of logit layer distillation being greatly overlooked. To provide a novel viewpoint for studying logits-based KD methods, we introduce the idea of decoupling into graph knowledge distillation. Specifically, we first reformulate the classical graph knowledge distillation loss into two parts, i.e., the target class graph distillation (TCGD) loss and the non-target class graph distillation (NCGD) loss. Next, we decouple the negative correlation between GNN's prediction confidence and NCGD loss, as well as eliminate the fixed weight between TCGD and NCGD. We named this logits-based method Decoupled Graph Knowledge Distillation (DGKD). It can flexibly adjust the weights of TCGD and NCGD for different data samples, thereby improving the prediction accuracy of the student MLP. Extensive experiments conducted on public benchmark datasets show the effectiveness of our method. Additionally, DGKD can be incorporated into any existing graph knowledge distillation framework as a plug-and-play loss function, further improving distillation performance. The code is available at https://github.com/xsk160/DGKD.

Neural Network-Based Prediction of Perceived Sleep Quality Through Wearable Device Data.

BACKGROUND: This study focuses on the development of a neural network model to predict perceived sleep quality using data from wearable devices. We collected various physiological metrics from 18 participants over four weeks, including heart rate, physical activity, and both device-measured and self-reported sleep quality. OBJECTIVES: The primary objective was to correlate wearable device data with subjective sleep quality perceptions. METHODS: Our approach used data processing, feature engineering, and optimizing a Multi-Layer Perceptron classifier. RESULTS: Despite comprehensive data analysis and model experimentation, the predictive accuracy for perceived sleep quality was moderate (59%), highlighting the complexities in accurately quantifying subjective sleep experiences through wearable data. Applying a tolerance of 1 grade (on a scale from 1-5), increased accuracy to 92%. DISCUSSION: More in-depth analysis is required to fully comprehend how wearables and artificial intelligence might assist in understanding sleep behavior.

Predicting Characteristics of Dissimilar Laser Welded Polymeric Joints Using a Multi-Layer Perceptrons Model Coupled with Archimedes Optimizer.

This study investigates the application of a coupled multi-layer perceptrons (MLP) model with Archimedes optimizer (AO) to predict characteristics of dissimilar lap joints made of polymethyl methacrylate (PMMA) and polycarbonate (PC). The joints were welded using the laser transmission welding (LTW) technique equipped with a beam wobbling feature. The inputs of the models were laser power, welding speed, pulse frequency, wobble frequency, and wobble width; whereas, the outputs were seam width and shear strength of the joint. The Archimedes optimizer was employed to obtain the optimal internal parameters of the multi-layer perceptrons. In addition to the Archimedes optimizer, the conventional gradient descent technique, as well as the particle swarm optimizer (PSO), was employed as internal optimizers of the multi-layer perceptrons model. The prediction accuracy of the three models was compared using different error measures. The AO-MLP outperformed the other two models. The computed root mean square errors of the MLP, PSO-MLP, and AO-MLP models are (39.798, 19.909, and 2.283) and (0.153, 0.084, and 0.0321) for shear strength and seam width, respectively.

Prediction of Oncotype DX recurrence score using deep multi-layer perceptrons in estrogen receptor-positive, HER2-negative breast cancer.

Oncotype DX (ODX) is a multi-gene expression signature designed for estrogen receptor (ER)-positive and human epidermal growth factor receptor 2 (HER2)-negative breast cancer patients to predict the recurrence score (RS) and chemotherapy (CT) benefit. The aim of our study is to develop a prediction tool for the three RS's categories based on deep multi-layer perceptrons (DMLP) and using only the morphoimmunohistological variables. We performed a retrospective cohort of 320 patients who underwent ODX testing from three French hospitals. Clinico-pathological characteristics were recorded. We built a supervised machine learning classification model using Matlab software with 152 cases for the training and 168 cases for the testing. Three classifiers were used to learn the three risk categories of the ODX, namely the low, intermediate, and high risk. Experimental results provide the area under the curve (AUC), respectively, for the three risk categories: 0.63 [95% confidence interval: (0.5446, 0.7154), p < 0.001], 0.59 [95% confidence interval: (0.5031, 0.6769), p < 0.001], 0.75 [95% confidence interval: (0.6184, 0.8816), p < 0.001]. Concordance rate between actual RS and predicted RS ranged from 53 to 56% for each class between DMLP and ODX. The concordance rate of low and intermediate combined risk group was 85%.We developed a predictive machine learning model that could help to define patient's RS. Moreover, we integrated histopathological data and DMLP results to select tumor for ODX testing. Thus, this process allows more relevant use of histopathological data, and optimizes and enhances this information.

A regression-tree multilayer-perceptron hybrid strategy for the prediction of ore crushing-plate lifetimes.

Highly tensile manganese steel is in great demand owing to its high tensile strength under shock loads. All workpieces are produced through casting, because it is highly difficult to machine. The probabilistic aspects of its casting, its variable composition, and the different casting techniques must all be considered for the optimisation of its mechanical properties. A hybrid strategy is therefore proposed which combines decision trees and artificial neural networks (ANNs) for accurate and reliable prediction models for ore crushing plate lifetimes. The strategic blend of these two high-accuracy prediction models is used to generate simple decision trees which can reveal the main dataset features, thereby facilitating decision-making. Following a complexity analysis of a dataset with 450 different plates, the best model consisted of 9 different multilayer perceptrons, the inputs of which were only the Fe and Mn plate compositions. The model recorded a low root mean square error (RMSE) of only 0.0614 h for the lifetime of the plate: a very accurate result considering their varied lifetimes of between 746 and 6902 h in the dataset. Finally, the use of these models under real industrial conditions is presented in a heat map, namely a 2D representation of the main manufacturing process inputs with a colour scale which shows the predicted output, i.e. the expected lifetime of the manufactured plates. Thus, the hybrid strategy extracts core training dataset information in high-accuracy prediction models. This novel strategy merges the different capabilities of two families of machine-learning algorithms. It provides a high-accuracy industrial tool for the prediction of the full lifetime of highly tensile manganese steel plates. The results yielded a precision prediction of (RMSE of 0.061 h) for the full lifetime of (light, medium, and heavy) crusher plates manufactured with the three (experimental, classic, and highly efficient (new)) casting methods.

Comparing Different Approaches for Mapping Urban Vegetation Cover from Landsat ETM+ Data: A Case Study on Brussels.

Urban growth and its related environmental problems call for sustainable urban management policies to safeguard the quality of urban environments. Vegetation plays an important part in this as it provides ecological, social, health and economic benefits to a city's inhabitants. Remotely sensed data are of great value to monitor urban green and despite the clear advantages of contemporary high resolution images, the benefits of medium resolution data should not be discarded. The objective of this research was to estimate fractional vegetation cover from a Landsat ETM+ image with sub-pixel classification, and to compare accuracies obtained with multiple stepwise regression analysis, linear spectral unmixing and multi-layer perceptrons (MLP) at the level of meaningful urban spatial entities. Despite the small, but nevertheless statistically significant differences at pixel level between the alternative approaches, the spatial pattern of vegetation cover and estimation errors is clearly distinctive at neighbourhood level. At this spatially aggregated level, a simple regression model appears to attain sufficient accuracy. For mapping at a spatially more detailed level, the MLP seems to be the most appropriate choice. Brightness normalisation only appeared to affect the linear models, especially the linear spectral unmixing.

Modeling Uncertainties in EEG Microstates: Analysis of Real and Imagined Motor Movements Using Probabilistic Clustering-Driven Training of Probabilistic Neural Networks.

Part of the process of EEG microstate estimation involves clustering EEG channel data at the global field power (GFP) maxima, very commonly using a modified K-means approach. Clustering has also been done deterministically, despite there being uncertainties in multiple stages of the microstate analysis, including the GFP peak definition, the clustering itself and in the post-clustering assignment of microstates back onto the EEG timecourse of interest. We perform a fully probabilistic microstate clustering and labeling, to account for these sources of uncertainty using the closest probabilistic analog to KM called Fuzzy C-means (FCM). We train softmax multi-layer perceptrons (MLPs) using the KM and FCM-inferred cluster assignments as target labels, to then allow for probabilistic labeling of the full EEG data instead of the usual correlation-based deterministic microstate label assignment typically used. We assess the merits of the probabilistic analysis vs. the deterministic approaches in EEG data recorded while participants perform real or imagined motor movements from a publicly available data set of 109 subjects. Though FCM group template maps that are almost topographically identical to KM were found, there is considerable uncertainty in the subsequent assignment of microstate labels. In general, imagined motor movements are less predictable on a time point-by-time point basis, possibly reflecting the more exploratory nature of the brain state during imagined, compared to during real motor movements. We find that some relationships may be more evident using FCM than using KM and propose that future microstate analysis should preferably be performed probabilistically rather than deterministically, especially in situations such as with brain computer interfaces, where both training and applying models of microstates need to account for uncertainty. Probabilistic neural network-driven microstate assignment has a number of advantages that we have discussed, which are likely to be further developed and exploited in future studies. In conclusion, probabilistic clustering and a probabilistic neural network-driven approach to microstate analysis is likely to better model and reveal details and the variability hidden in current deterministic and binarized microstate assignment and analyses.