Machine Learning Optimization Research Articles

Crafting a Personalized Prognostic Model for Malignant Prostate Cancer Patients Using Risk Gene Signatures Discovered through TCGA-PRAD Mining, Machine Learning, and Single-Cell RNA-Sequencing.

Prostate cancer is a significant clinical issue, particularly for high Gleason score (GS) malignancy patients. Our study aimed to engineer and validate a risk model based on the profiles of high-GS PCa patients for early identification and the prediction of prognosis. We conducted differential gene expression analysis on patient samples from The Cancer Genome Atlas (TCGA) and enriched our understanding of gene functions. Using the least absolute selection and shrinkage operator (LASSO) regression, we established a risk model and validated it using an independent dataset from the International Cancer Genome Consortium (ICGC). Clinical variables were incorporated into a nomogram to predict overall survival (OS), and machine learning was used to explore the risk factor characteristics' impact on PCa prognosis. Our prognostic model was confirmed using various databases, including single-cell RNA-sequencing datasets (scRNA-seq), the Cancer Cell Line Encyclopedia (CCLE), PCa cell lines, and tumor tissues. We identified 83 differentially expressed genes (DEGs). Furthermore, WASIR1, KRTAP5-1, TLX1, KIF4A, and IQGAP3 were determined to be significant risk factors for OS and progression-free survival (PFS). Based on these five risk factors, we developed a risk model and nomogram for predicting OS and PFS, with a C-index of 0.823 (95% CI, 0.766-0.881) and a 10-year area under the curve (AUC) value of 0.788 (95% CI, 0.633-0.943). Additionally, the 3-year AUC was 0.759 when validating using ICGC. KRTAP5-1 and WASIR1 were found to be the most influential prognosis factors when using the optimized machine learning model. Finally, the established model was interrelated with immune cell infiltration, and the signals were found to be differentially expressed in PCa cells when using scRNA-seq datasets and tissues. We engineered an original and novel prognostic model based on five gene signatures through TCGA and machine learning, providing new insights into the risk of scarification and survival prediction for PCa patients in clinical practice.

Technical Perspective: Conjunctive Queries with Comparisons

Query processing, the art of efficiently executing a relational query on a given database, is a foundational and core area in data management research. Established at the dawn of relational database systems in the 1970's, relational query processing remains a highly relevant and vibrant research topic today as recent work shows that, apart from its application in traditional database scenarios, it is also highly effective in optimizing machine learning workloads [1].

Performance of evolutionary optimized machine learning for modeling total organic carbon in core samples of shale gas fields

Rock samples' TOC content is the best indicator of the organic matter in source rocks. The origin rock samples’ analysis is used to calculate it manually by specialists. This method requires time and resources because it relies on samples from many well intervals in source rocks. Therefore, research has been done to aid this effort. Machine learning algorithms can estimate total organic carbon instead of well logs and stratigraphic studies. In light of these efforts, the current work present a study on automating the total organic carbon estimation using machine learning approaches improved by an evolutionary methodology to give the model flexibility and precision. Genetic algorithms, differential evolution, particle swarm optimization, grey wolf optimization, artificial bee colony, and evolution strategies were used to improve machine learning models to predict TOC. The six metaheuristics were integrated into four machine learning methods: extreme learning machine, elastic net linear model, linear support vector regression, and multivariate adaptive regression splines. Core samples from the YuDong-Nan shale gas field, located in the Sichuan basin, were used to evaluate the hybrid strategy. The findings show that combining machine learning models with an evolutionary algorithms in a hybrid fashion produce flexible models that accurately predict TOC. The results show that, independent of the metaheuristic used to guide the model selection, optimized extreme learning machines attained the best performance scores according to six metrics. Such hybrid models can be used in exploratory geological research, particularly for unconventional oil and gas resources.

RWD145 Machine Learning and Artificial Intelligence for Clinical Trial Optimization: A Review of Opportunities to Leverage Real World Data

Clinical trials for novel therapeutics are expensive, time-consuming, and have an end-to-end success rate of under 8%. Several applications of artificial intelligence and machine learning (AI/ML) to trial design and conduct have the potential to increase trial efficiency, and the use of real world data (RWD) is often critical to the development of these novel applications. Objectives: The objective of this study was to comprehensively review applications of AI/ML that used RWD to improve clinical trial design and conduct.

Representation Learning-Driven Fully Automated Framework for the Inverse Design of Frequency-Selective Surfaces

Frequency-selective surfaces (FSSs) refer to planar structures that behave with specific electromagnetic (EM) responses within a frequency range and are widely applied in wireless propagation systems. Given the fact that different EM responses correspond to distinguished topologies, conventional inverse design methods of FSSs are usually labor-intensive, as they rely on experienced human engineers to determine the topology and then rationally tune its structures. There have been great attempts using optimization algorithms (e.g., genetic algorithms) or machine learning to automate the second tuning stage after the initial EM topologies are determined by human engineers. However, the first topology selection stage still requires engagements with experienced engineers. This article proposes a fully automated framework for the inverse design of FSSs. We achieved a fully automated inverse design by establishing a machine-friendly mapping flow. The mapping flow derives its continuity and compactness from representation learning, which enables both autoselection of the topology and autoevolution of the unit cell based on the topology. The autoselection stage automatically determines the appropriate topology by compressing the EM constraints through the principal component analysis (PCA) and classifying the topology using the support vector machine (SVM). Afterward, the autoevolution system can efficiently evolve until it yields an optimal unit cell. We developed a self-monitor strategy to control the evolution and maximize the evolution efficiency by adaptively tuning the three modules within the autoevolution system. We validated the presented framework with four FSS designs. The results proved its potential as a highly efficient fully automated tool for the inverse design of FSSs.

A Simplified Constrained Bayesian Optimization Approach to Optimize the Tx Equalization in SerDes Channels

Design of high-speed channels has become increasingly more complicated. Due to the eye diagram closure at higher speeds, designers use Tx equalization by placing an FIR filter at Tx. Assigning the FIR tap values can be time consuming and require domain expertise since it can require sweeping hundreds or more combinations of tap values. Therefore, in this paper we propose a machine learning optimization approach to find the FIR tap values which result in the largest eye opening. Conventional optimization techniques may not be applicable in this context since specifications of the channel can require a constraint which is sum of the absolute value of the FIR taps needs to be equal to 1. Therefore, we have developed a simplified constrained Bayesian optimization approach that can automate this process and expedite calculation of the FIR tap values without requiring domain expertise. Numerical examples are provided to show efficiency of the proposed approach and compare its performance with Bayesian optimization and Genetic algorithm for this problem.

Assessing the predictive capacity of machine learning models using patient-specific variables in determining in-hospital outcomes after THA

BackgroundMachine learning is a subset of artificial intelligence using algorithmic modeling to progressively learn and create predictive models. Clinical application of machine learning can aid physicians through identification of risk factors and implications of predicted patient outcomes. AimsThe aim of this study was to compare patient-specific and situation perioperative variables through optimized machine learning models to predict postoperative outcomes. MethodsData from 2016 to 2017 from the National Inpatient Sample was used to identify 177,442 discharges undergoing primary total hip arthroplasty, which were included in the training, testing, and validation of 10 machine learning models. 15 predictive variables consisting of 8 patient-specific and 7 situational specific variables were utilized to predict 3 outcome variables: length of stay, discharge, and mortality. The machine learning models were assessed in responsiveness via area under the curve and reliability. ResultsFor all outcomes, Linear Support Vector Machine had the highest responsiveness among all models when using all variables. When utilizing patient-specific variables only, responsiveness of the top 3 models ranged between 0.639 and 0.717 for length of stay, 0.703–0.786 for discharge disposition, and 0.887–0.952 for mortality. The top 3 models utilizing situational variables only produced responsiveness of 0.552–0.589 for length of stay, 0.543–0.574 for discharge disposition, and 0.469–0.536 for mortality. ConclusionsLinear Support Vector Machine was the most responsive machine learning model of the 10 algorithms trained, while decision list was most reliable. Responsiveness was observed to be consistently higher with patient-specific variables than situational variables, emphasizing the predictive capacity and value of patient-specific variables. The current practice in machine learning literature generally deploys a single model, it is suboptimal to develop optimized models for application into clinical practice. The limitation of other algorithms may prohibit potential more reliable and responsive models.Level of Evidence III.

Data-driven optimization of FePt heat-assisted magnetic recording media accelerated by deep learning TEM image segmentation

The main bottleneck for heat-assisted magnetic recording (HAMR) to achieve a potential areal density of 4 Tb/in2 is the difficulty in obtaining FePt-X nanogranular media with an ideal stacking structure of perfectly isolated L10-FePt columnar nanograins. Here, we present a fully automated routine that combines a convolutional neural network and machine vision to enable data mining from transmission electron microscopy images of FePt-C nanogranular media. This allowed us to generate a dataset and implement a machine learning optimization model that guides process parameters to achieve the desired nanostructure, i.e., small grain size with unimodal distribution and a large coercivity, which was successfully validated experimentally. This work demonstrates the promise of data-driven design of high-density HAMR media.

Power Transformers Health Index Enhancement Based on Convolutional Neural Network after Applying Imbalanced-Data Oversampling

The transformer health index (HI) concept has been used as an important part of management resources and is implemented for the state assessment and ranking of Power transformers. The HI state is estimated based on many power transformer oil parameters. However, the main problem in the HI procedure as a diagnostic method is the presence of routine measurements and accurate test results. The power transformer HI prediction is carried out in this work using 1361 dataset samples collected from two different utilities. The proposed model is used to predict and diagnose the HI state of the power transformer by using a convolutional neural network (CNN) approach. The imbalance between the training dataset sample classes produces a good prediction of the class with a major number of samples while a low detection of the class has a minor number of samples. The oversampling approach is used to balance the training samples to enhance the prediction accuracy of the classification methods. The proposed CNN model predicts the HI of the power transformers after applying the oversampling approach to the training dataset samples. The results obtained with the proposed CNN model are compared with those obtained with the optimized machine learning (ML) classification methods with the superiority of the CNN results. Feature reductions are applied to minimize testing time, effort, and costs. Finally, the proposed CNN model is checked with uncertain noise in full and reduced features of up to ±25% with a good prediction diagnosis of the power transformer HI.

Engineering Applications of Artificial Intelligence in Mechanical Design and Optimization

This study offers a complete analysis of the use of deep learning or machine learning, as well as precise recommendations on how these methods could be used in the creation of machine components and nodes. The examples in this thesis are intended to identify areas in mechanical design and optimization where this technique could be widely applied in the future, benefiting society and advancing the current state of modern mechanical engineering. The review begins with a discussion on the workings of artificial intelligence, machine learning, and deep learning. Different techniques, classifications, and even comparisons of each method are described in detail. The most common programming languages, frameworks, and software used in mechanical engineering for this problem are gradually introduced. Input data formats and the most common datasets that are suitable for the field of machine learning in mechanical design and optimization are also discussed. The second half of the review describes the current use of machine learning in several areas of mechanical design and optimization, using specific examples that have been investigated by researchers from around the world. Further research directions on the use of machine learning and neural networks in the fields of mechanical design and optimization are discussed.

An aggressive reduction on the complexity of optimization for non-strongly convex objectives

Tremendous efficient optimization methods have been proposed for strongly convex objectives optimization in modern machine learning. For non-strongly convex objectives, a popular approach is to apply a reduction from non-strongly convex to a strongly convex case via regularization techniques. Reduction on objectives with adaptive decrease on regularization tightens the optimal convergence of algorithms to be independent on logarithm factor. However, the initialization of parameter of regularization has a great impact on the performance of the reduction. In this paper, we propose an aggressive reduction to reduce the complexity of optimization for non-strongly convex objectives, and our reduction eliminates the impact of the initialization of parameter on the convergent performances of algorithms. Aggressive reduction not only adaptively decreases the regularization parameter, but also modifies regularization term as the distance between current point and the approximate minimizer. Our aggressive reduction can also shave off the non-optimal logarithm term theoretically, and make the convergent performance of algorithm more compact practically. Experimental results on logistic regression and image deblurring confirm this success in practice.

Reconstruction of electron radiation belts using data assimilation and machine learning

We present a reconstruction of radiation belt electron fluxes using data assimilation with low-Earth-orbiting Polar Orbiting Environmental Satellites (POES) measurements mapped to near equatorial regions. Such mapping is a challenging task and the appropriate methodology should be selected. To map POES measurements, we explore two machine learning methods: multivariate linear regression (MLR) and neural network (NN). The reconstructed flux is included in data assimilation with the Versatile Electron Radiation Belts (VERB) model and compared with Van Allen Probes and GOES observations. We demonstrate that data assimilation using MLR-based mapping provides a reasonably good agreement with observations. Furthermore, the data assimilation with the flux reconstructed by NN provides better performance in comparison to the data assimilation using flux reconstructed by MLR. However, the improvement by adding data assimilation is limited when compared to the purely NN model which by itself already has a high performance of predicting electron fluxes at high altitudes. In the case an optimized machine learning model is not possible, our results suggest that data assimilation can be beneficial for reconstructing outer belt electrons by correcting errors of a machine learning based LEO-to-MEO mapping and by providing physics-based extrapolation to the parameter space portion not included in the LEO-to-MEO mapping, such as at the GEO orbit in this study.

Assessment of long-term mangrove distribution using optimised machine learning algorithms and landscape pattern analysis.

Mangrove ecosystems provide numerous benefits, including carbon storage, coastal protection and food for marine organisms. However, mapping and monitoring of mangrove status in some regions, such as the Red Sea area, has been hindered by a lack of data, accurate and precise maps and technical expertise. In this study, an advanced machine learning algorithm was proposed to produce an accurate and precise high-resolution land use map that includes mangroves in the Al Wajh Bank habitat in northeastern Saudi Arabia. To achieve this, high-resolution multispectral images were generated using an image fusion technique, and machine learning algorithms were applied, including artificial neural networks, random forests and support vector machine algorithms. The performance of the models was evaluated using various matrices, and changes in mangrove distribution and connectivity were assessed using the landscape fragmentation model and Getis-Ord statistics. The research gap that this study aims to address is the lack of accurate and precise mapping and assessment of mangrove status in the Red Sea area, particularly in data-scarce regions. Our study produced high-resolution mobile laser scanning (MLS) imagery of 15-m length for 2014 and 2022, and trained 5, 6 and 9 models for artificial neural networks, support vector machines and random forests (RF) to predict land use and land cover maps using 15-m and 30-m resolution MLS images. The best models were identified using error matrices, and it was found that RF outperformed other models. According to the 15-m resolution map of 2022 and the best models of RF, the mangrove cover in the Al Wajh Bank is 27.6 km2, which increased to 34.99 km2 in the case of the 30-m resolution image of 2022, and was 11.94 km2 in 2014, indicating a doubling of the mangrove area. Landscape structure analysis revealed an increase in small core and hotspot areas, which were converted into medium core and very large hotspot areas in 2014. New mangrove areas were identified in the form of patches, edges, potholes and coldspots. The connectivity model showed an increase in connectivity over time, promoting biodiversity. Our study contributes to the promotion of the protection, conservation and planting of mangroves in the Red Sea area.

4E analysis and machine learning optimization of a geothermal-based system integrated with ejector refrigeration cycle for efficient hydrogen production and liquefaction

This study presents a new geothermal-based system for power generation and liquid hydrogen production. Depending on the network requirements, part of the generated power is supplied to the network and the remaining part enters the PEM electrolyzer to produce hydrogen. The produced hydrogen is then pre-cooled by the ejector cooling cycle before entering the liquefaction cycle. The ejector refrigeration cycle not only uses geothermal waste heat but also facilitates the liquefaction process through precooling hydrogen; therefore, it significantly improves the overall performance of the integrated system. In addition, the modified liquefaction cycle is suggested, which is superior to common ones. After simulating, some key parameters are identified and a comprehensive parametric research is carried out to investigate the thermodynamic, economic and environmental performance in different conditions. The optimization procedure is implemented through a genetic algorithm approach and the objective functions are defined. The results reveal that the double flash power cycle has the greatest effect on exergy destruction rate and total cost rate. In the optimal case, the total cost rate, the liquefied hydrogen rate, the total exergy efficiency, net power generation, and CO2 reduction rate are determined as 181.71 $/h, 59.92 kg/h, 25.27%, 4.03 MW, and 1421.2 kg/h, respectively. Also, coupling an artificial neural network with genetic algorithm significantly lessens optimization time.

Multi‐Objective Optimization During Magnetic Field Assisted ECDM of Alumina‐Zirconium Dioxide Ceramic Composite using Battle Royale Optimization

Alumina‐Zirconium dioxide (Al2O3‐ZrO2) Ceramic Composite Material (CCM) is specifically known for its enhanced mechanical and corrosion resistance properties and is widely used as raw material for mechanical parts like, pump components, die inserts, bearings, etc. As a result, industrialists are searching for an efficient method for machining this Al2O3‐ZrO2 material. In this regard, a hybrid unconventional machining process called Electrochemical Discharge Machining (ECDM) is adapted to analyze the machinability of Al2O3‐ZrO2 CCM. Besides, to ensure the efficiency of the ECDM process, a magnetic field is also given to the tool electrode during this study to improve the Material Removal Rate (MRR) of the ECDM machine. The experiment is designed using Response Surface Methodology (RSM) by changing the magnitudes of input controls, namely Electrolytic Concentration (EC), Inter‐electrode Gap (IEG) and Applied Voltage (AV). Moreover, a novel hybrid machine learning optimization strategy called Deep Belief Network based Battle Royal Optimization (DBN‐BRO) algorithm is developed to predict and optimize the ECDM process. Finally, the optimum results are perceived from 55 V AV, 22.727% EC and 40.909 mm IEG input levels. The proposed method shows less than 0.6207 Root Mean Square Error (RMSE) and tools nearly 80 iterations for optimizing the results.

GaNDLF: the generally nuanced deep learning framework for scalable end-to-end clinical workflows

Deep Learning (DL) has the potential to optimize machine learning in both the scientific and clinical communities. However, greater expertise is required to develop DL algorithms, and the variability of implementations hinders their reproducibility, translation, and deployment. Here we present the community-driven Generally Nuanced Deep Learning Framework (GaNDLF), with the goal of lowering these barriers. GaNDLF makes the mechanism of DL development, training, and inference more stable, reproducible, interpretable, and scalable, without requiring an extensive technical background. GaNDLF aims to provide an end-to-end solution for all DL-related tasks in computational precision medicine. We demonstrate the ability of GaNDLF to analyze both radiology and histology images, with built-in support for k-fold cross-validation, data augmentation, multiple modalities and output classes. Our quantitative performance evaluation on numerous use cases, anatomies, and computational tasks supports GaNDLF as a robust application framework for deployment in clinical workflows.

A Model-Based Approach Machine Learning to Scalable Portfolio Selection

This study proposes a scalable asset selection and allocation approach using machine learning that integrates clustering methods into portfolio optimization models. The methodology applies the Uniform Manifold Approximation and Projection method and ensemble clustering techniques to preselect assets from the Ibovespa and S&P 500 indices. The research compares three allocation models and finds that the Hierarchical Risk Parity model outperformed the others, with a Sharpe ratio of 1.11. Despite the pandemic's impact on the portfolios, with drawdowns close to 30%, they recovered in 111 to 149 trading days. The portfolios outperformed the indices in cumulative returns, with similar annual volatilities of 20%. Preprocessing with UMAP allowed for finding clusters with higher discriminatory power, evaluated through internal cluster validation metrics, helping to reduce the problem's size during optimal portfolio allocation. Overall, this study highlights the potential of machine learning in portfolio optimization, providing a useful framework for investment practitioners.

Optimizing the input feature sets and machine learning algorithms for reliable and accurate estimation of continuous, cuffless blood pressure

The advent of mobile devices, wearables and digital healthcare has unleashed a demand for accurate, reliable, and non-interventional ways to measure continuous blood pressure (BP). Many consumer products claim to measure BP with a cuffless device, but their lack of accuracy and reliability limit clinical adoption. Here, we demonstrate how multimodal feature datasets, comprising: (i) pulse arrival time (PAT); (ii) pulse wave morphology (PWM), and (iii) demographic data, can be combined with optimized Machine Learning (ML) algorithms to estimate Systolic BP (SBP), Diastolic BP (DBP) and Mean Arterial Pressure (MAP) within a 5 mmHg bias of the gold standard Intra-Arterial BP, well within the acceptable limits of the IEC/ANSI 80601-2-30 (2018) standard. Furthermore, DBP’s calculated using 126 datasets collected from 31 hemodynamically compromised patients had a standard deviation within 8 mmHg, while SBP’s and MAP’s exceeded these limits. Using ANOVA and Levene’s test for error means and standard deviations, we found significant differences in the various ML algorithms but found no significant differences amongst the multimodal feature datasets. Optimized ML algorithms and key multimodal features obtained from larger real-world data (RWD) sets could enable more reliable and accurate estimation of continuous BP in cuffless devices, accelerating wider clinical adoption.

Proficient job scheduling in cloud computation using an optimized machine learning strategy

Proficient job scheduling in cloud computation using an optimized machine learning strategy

A CT-based radiomics nomogram for distinguishing between malignant and benign Bosniak IIF masses: a two-centre study

To establish and assess a computed tomography (CT)-based radiomics nomogram for identifying malignant and benign Bosniak IIF masses. In total, 150 patients with Bosniak IIF masses were separated into a training set (n=106) and a test set (n=44) in a ratio of 7:3. A radiomics signature was calculated based on extracted features from the three phases of CT images. A clinical model was constructed based on clinical characteristics and CT features, and a nomogram incorporating the radiomics signature and independent clinical variables was established. The calibration ability, discrimination accuracy, and clinical value of the nomogram model were assessed. Twelve features derived from CT images were applied to establish the radiomics signature. The performance levels of three machine-learning models were improved by adding the synthetic minority oversampling technique algorithm. The optimised machine learning model was a combination of the minimum redundancy maximum relevance-least absolute shrinkage and selection operator feature screening method+logistic regression classifier+synthetic minority oversampling technique algorithm, which demonstrated excellent identification ability on the test set (area under the curve [AUC], 0.970; 95% confidence interval [CI], 0.940-1.000). The nomogram model displayed outstanding discrimination ability on the test set (AUC, 0.972; 95% CI, 0.942-1.000). The CT-based radiomics nomogram was useful for discriminating between malignant and benign Bosniak IIF masses, which improved the precision of preoperative diagnosis.