Open Source Machine Learning Research Articles

Exploring the Efficiency of Hybrid Recommender Systems Implemented with TensorFlow Framework

In recent years, the field of recommendation systems has seen significant advancement with the introduction of hybrid approaches. These systems combine the strengths of multiple recommendation techniques to provide more accurate and diverse recommendations to users. In this research, we propose and evaluate the effectiveness of a hybrid recommender system that utilizes TensorFlow, an open-source machine learning framework, to implement the system. The proposed system combines both collaborative and content-based methods to remove the cold start problem and make personalized recommendations that can recommend similar movies to the users based on features extracted by the model. The results of our experiments demonstrate that the proposed hybrid system outperforms traditional singular methods and can be effectively implemented using TensorFlow. This research provides insights into the potential of TensorFlow for building efficient hybrid recommendation systems and the benefits of combining multiple recommendation techniques.

Bridging the language gap: an empirical study of bindings for open source machine learning libraries across software package ecosystems

Bridging the language gap: an empirical study of bindings for open source machine learning libraries across software package ecosystems

Tailoring Risk Prediction Models to Local Populations.

Risk estimation is an integral part of cardiovascular care. Local recalibration of guideline-recommended models could address the limitations of existing tools. To provide a machine learning (ML) approach to augment the performance of the American Heart Association's Predicting Risk of Cardiovascular Disease Events (AHA-PREVENT) equations when applied to a local population while preserving clinical interpretability. This cohort study used a New England-based electronic health record cohort of patients without prior atherosclerotic cardiovascular disease (ASCVD) who had the data necessary to calculate the AHA-PREVENT 10-year risk of developing ASCVD in the event period (2007-2016). Patients with prior ASCVD events, death prior to 2007, or age 79 years or older in 2007 were subsequently excluded. The final study population of 95 326 patients was split into 3 nonoverlapping subsets for training, testing, and validation. The AHA-PREVENT model was adapted to this local population using the open-source ML model (MLM) Extreme Gradient Boosting model (XGBoost) with minimal predictor variables, including age, sex, and AHA-PREVENT. The MLM was monotonically constrained to preserve known associations between risk factors and ASCVD risk. Along with sex, race and ethnicity data from the electronic health record were collected to validate the performance of ASCVD risk prediction in subgroups. Data were analyzed from August 2021 to February 2024. Consistent with the AHA-PREVENT model, ASCVD events were defined as the first occurrence of either nonfatal myocardial infarction, coronary artery disease, ischemic stroke, or cardiovascular death. Cardiovascular death was coded via government registries. Discrimination, calibration, and risk reclassification were assessed using the Harrell C index, a modified Hosmer-Lemeshow goodness-of-fit test and calibration curves, and reclassification tables, respectively. In the test set of 38 137 patients (mean [SD] age, 64.8 [6.9] years, 22 708 [59.5]% women and 15 429 [40.5%] men; 935 [2.5%] Asian, 2153 [5.6%] Black, 1414 [3.7%] Hispanic, 31 400 [82.3%] White, and 2235 [5.9%] other, including American Indian, multiple races, unspecified, and unrecorded, consolidated owing to small numbers), MLM-PREVENT had improved calibration (modified Hosmer-Lemeshow P > .05) compared to the AHA-PREVENT model across risk categories in the overall cohort (χ23 = 2.2; P = .53 vs χ23 > 16.3; P < .001) and sex subgroups (men: χ23 = 2.1; P = .55 vs χ23 > 16.3; P < .001; women: χ23 = 6.5; P = .09 vs. χ23 > 16.3; P < .001), while also surpassing a traditional recalibration approach. MLM-PREVENT maintained or improved AHA-PREVENT's calibration in Asian, Black, and White individuals. Both MLM-PREVENT and AHA-PREVENT performed equally well in discriminating risk (approximate ΔC index, ±0.01). Using a clinically significant 7.5% risk threshold, MLM-PREVENT reclassified a total of 11.5% of patients. We visualize the recalibration through MLM-PREVENT ASCVD risk charts that highlight preserved risk associations of the original AHA-PREVENT model. The interpretable ML approach presented in this article enhanced the accuracy of the AHA-PREVENT model when applied to a local population while still preserving the risk associations found by the original model. This method has the potential to recalibrate other established risk tools and is implementable in electronic health record systems for improved cardiovascular risk assessment.

Computer vision for kinematic metrics of the drinking task in a pilot study of neurotypical participants

Assessment of the upper limb is critical to guiding the rehabilitation cycle. Drawbacks of observation-based assessment include subjectivity and coarse resolution of ordinal scales. Kinematic assessment gives rise to objective quantitative metrics, but uptake is encumbered by costly and impractical setups. Our objective was to investigate feasibility and accuracy of computer vision (CV) for acquiring kinematic metrics of the drinking task, which are recommended in stroke rehabilitation research. We implemented CV for upper limb kinematic assessment using modest cameras and an open-source machine learning solution. To explore feasibility, 10 neurotypical participants were recruited for repeated kinematic measures during the drinking task. To investigate accuracy, a simultaneous marker-based motion capture system was used, and error was quantified for the following kinematic metrics: Number of Movement Units (NMU), Trunk Displacement (TD), and Movement Time (MT). Across all participant trials, kinematic metrics of the drinking task were successfully acquired using CV. Compared to marker-based motion capture, no significant difference was observed for group mean values of kinematic metrics. Mean error for NMU, TD, and MT were − 0.12 units, 3.4 mm, and 0.15 s, respectively. Bland–Altman analysis revealed no bias. Kinematic metrics of the drinking task can be measured using CV, and preliminary findings support accuracy. Further study in neurodivergent populations is needed to determine validity of CV for kinematic assessment of the post-stroke upper limb.

Using matrix assisted laser desorption ionisation mass spectrometry combined with machine learning for vaccine authenticity screening

The global population is increasingly reliant on vaccines to maintain population health with billions of doses used annually in immunisation programmes. Substandard and falsified vaccines are becoming more prevalent, caused by both the degradation of authentic vaccines but also deliberately falsified vaccine products. These threaten public health, and the increase in vaccine falsification is now a major concern. There is currently no coordinated global infrastructure or screening methods to monitor vaccine supply chains. In this study, we developed and validated a matrix-assisted laser desorption/ionisation-mass spectrometry (MALDI-MS) workflow that used open-source machine learning and statistical analysis to distinguish authentic and falsified vaccines. We validated the method on two different MALDI-MS instruments used worldwide for clinical applications. Our results show that multivariate data modelling and diagnostic mass spectra can be used to distinguish authentic and falsified vaccines providing proof-of-concept that MALDI-MS can be used as a screening tool to monitor vaccine supply chains.

Satellite Image Classification using TensorFlow

Abstract: Satellite image classification plays a crucial role in various fields such as agriculture, urban planning, disaster management, and environmental monitoring. This paper presents a novel approach utilizing TensorFlow, a popular open-source machine learning framework, for satellite image classification. The proposed methodology leverages deep learning techniques to extract meaningful features from satellite images, enabling accurate classification into predefined categories. By harnessing the power of convolutional neural networks (CNNs) implemented in TensorFlow, this research aims to enhance the efficiency and accuracy of satellite image classification tasks. The experimental results demonstrate the effectiveness of the proposed approach in achieving high classification accuracy of about 96.5% while maintaining computational efficiency. It classifies about 10 different classes obtained from the EuroSat dataset. It is implemented in Google Collab for faster training and implementation.

Evaluating Molecular Complexity with Open-Source Machine Learning Approaches to Predict Process Mass Intensity.

The application of green chemistry is critical for cultivating environmental responsibility and sustainable practices in pharmaceutical manufacturing. Process mass intensity (PMI) is a key metric that quantifies the resource efficiency of a manufacturing process, but determining what constitutes a successful PMI of a specific molecule is challenging. A recent approach correlated molecular features to a crowdsourced definition of molecular complexity to determine PMI targets. While recent machine learning tools show promise in predicting molecular complexity, a more extensive application could significantly optimize manufacturing processes. To this end, we refine and expand upon the SMART-PMI tool by Sheridan et al. to create an open-source model and application. Our solution emphasizes explainability and parsimony to facilitate a nuanced understanding of prediction and ensure informed decision-making. The resulting model uses four descriptors-the heteroatom count, stereocenter count, unique topological torsion, and connectivity index chi4n-to compute molecular complexity with a comparable 82.6% predictive accuracy and 0.349 RMSE. We develop a corresponding app that takes in structured data files (SDF) to rapidly quantify molecular complexity and provide a PMI target that can be used to drive process development activities. By integrating machine learning explainability and open-source accessibility, we provide flexible tools to advance the field of green chemistry and sustainable pharmaceutical manufacturing.

Opensource machine learning metamodels for assessing blade performance impairment due to general leading edge degradation

Blades leading edge erosion can significantly reduce annual energy production of wind turbines. Accurate estimates of the resulting blade performance impairment are paramount to predict the resulting energy losses and enable cost-informed decisions on optimal maintenance and operational strategies, maximizing energy production and reducing maintenance costs. Computational Fluid Dynamics (CFD) is a robust approach for predicting the performance losses due to LEE. However, the impact of the damage on blade aerodynamics varies depending on damage pattern, extent and location. Therefore, direct CFD simulation of a sufficiently general set of damaged blades is computationally not viable in industrial applications, since the energy loss assessment needs to be performed for hundreds of turbines at many times of the wind farm operation. To address this issue, previous studies showed how CFD can be used to train machine learning metamodels of the perfomance of damaged blade sections, enabling the definition of multi-fidelity energy loss prediction systems. This study presents improved metamodels, using validated CFD to generate training datasets that cover a more general and wider range of erosion patterns, from low-amplitude roughness to severe grooves. In order to provide the industry with additional erosion geometry-linked tools for estimating energy yield losses, and foster further research and development in this area, the developed meta-models have been made available online with unrestricted access.

Comparative analysis of real issues in open-source machine learning projects

ContextIn the last decade of data-driven decision-making, Machine Learning (ML) systems reign supreme. Because of the different characteristics between ML and traditional Software Engineering systems, we do not know to what extent the issue-reporting needs are different, and to what extent these differences impact the issue resolution process.ObjectiveWe aim to compare the differences between ML and non-ML issues in open-source applied AI projects in terms of resolution time and size of fix. This research aims to enhance the predictability of maintenance tasks by providing valuable insights for issue reporting and task scheduling activities.MethodWe collect issue reports from Github repositories of open-source ML projects using an automatic approach, filter them using ML keywords and libraries, manually categorize them using an adapted deep learning bug taxonomy, and compare resolution time and fix size for ML and non-ML issues in a controlled sample.Result147 ML issues and 147 non-ML issues are collected for analysis. We found that ML issues take more time to resolve than non-ML issues, the median difference is 14 days. There is no significant difference in terms of size of fix between ML and non-ML issues. No significant differences are found between different ML issue categories in terms of resolution time and size of fix.ConclusionOur study provided evidence that the life cycle for ML issues is stretched, and thus further work is required to identify the reason. The results also highlighted the need for future work to design custom tooling to support faster resolution of ML issues.

Open-source machine learning BANTER acoustic classification of beaked whale echolocation pulses

Passive acoustic monitoring is increasingly used for assessing populations of marine mammals; however, analysis of large datasets is limited by our ability to easily classify sounds detected. Classification of beaked whale acoustic events, in particular, requires evaluation of multiple lines of evidence by expert analysts. Here we present a highly automated approach to acoustic detection and classification using supervised machine learning and open source software methods. Data from four large scale surveys of beaked whales (northwestern North Atlantic, southwestern North Atlantic, Hawaii, and eastern North Pacific) were analyzed using PAMGuard (acoustic detection), PAMpal (acoustic analysis) and BANTER (hierarchical random forest classifier). Overall classification accuracy ranged from 88% for the southwestern North Atlantic data to 97% for the northwestern North Atlantic. Results for many species could likely be improved with increased sample sizes, consideration of alternative automated detectors, and addition of relevant environmental features. These methods provide a highly automated approach to acoustic detection and classification using open source methods that can be readily adopted for other species and geographic regions.

Predictive Modeling Through Fusion of Passengers Information Transferred to Alternate Dimensions

This research focuses on the identification of passengers, in dimensions using information fusion as a tool. We recognize the challenges involved in identifying individuals who have been transferred to alternate dimensions and in this study we make use of CatBoost, an open source machine learning algorithm to address this problem. Our approach includes a preprocessing strategy that involves filling in missing values using techniques like priori distribution terms, which helps ensure the reliability of our dataset. By leveraging CatBoosts ability to handle variables and prevent overfitting we achieve results in accurately predicting passenger movement across dimensions. Our analysis highlights CatBoosts effectiveness in identifying patterns within data leading to more precise predictions for interdimensional passenger transportation. Additionally we incorporate techniques, like Greedy TS augmentation to enhance the adaptability of the algorithm and improve precision while reducing bias in modeling. Proof-of-concept experiments demonstrate that the proposed fusion system not only advances predictive modeling in niche domains but also paves the way for broader applications of machine learning in deciphering complex phenomena beyond traditional realms, marking a significant stride in understanding and addressing unconventional challenges.

Demystifying the Impact of Open-Source Machine Learning Libraries on Software Analytics

Demystifying the Impact of Open-Source Machine Learning Libraries on Software Analytics

Utility of Machine Learning in the Prediction of Post-Hepatectomy Liver Failure in Liver Cancer.

Posthepatectomy liver failure (PHLF) is a serious complication associated with high mortality rates. Machine learning (ML) has rapidly developed and may outperform traditional models in predicting PHLF in patients who have undergone hepatectomy. This study aimed to predict PHLF using ML and compare its performance with that of traditional scoring systems. The clinicopathological data of 334 patients who underwent liver resection were retrospectively collected. The Pycaret library, a simple, open-source machine learning library, was used to compare multiple classification models for PHLF prediction. The predictive performance of 15 ML algorithms was compared using the mean area under the receiver operating characteristic curve (AUROC) and accuracy, and the best-fit model was selected among 15 ML algorithms. Next, the predictive performance of the selected ML-PHLF model was compared with that of routine scoring systems, the albumin-bilirubin score (ALBI) and the fibrosis-4 (FIB-4) index, using AUROC. The best model was extreme gradient boosting (accuracy:93.1%; AUROC:0.863) among the 15 ML algorithms. As compared with ALBI and FIB-4, the ML PHLF model had higher AUROC for predicting PHLF. The novel ML model for predicting PHLF outperformed routine scoring systems.

Predicting Subcontractor Performance Using Artificial Intelligence

Subcontractors contribute to almost 90% of the overall construction work hence, assessing the performance of their work from commencement till the completion stage is essential. This paper focuses on identifying different factors affecting the subcontracted work performance and developing a predictive model using classification - based algorithm to find the proficient subcontractors.&#x0D; Data collected from the building construction projects was analyzed and utilized. Expert validation method was carried out to validate the factors that were obtained from literature review and a survey was conducted to assess the subcontractor’s performance level. Different classification algorithms such as Naïve Bayes, Logistic, Multilayer Perceptron, Sequential Minimal Optimization (SMO), KStar, J48 and Random Forest were applied to the collected data. Waikato Environment for Knowledge Analysis (WEKA), an opensource machine learning tool was used to compare the performance of several algorithms. Statics were generated and compared using k-folds cross validation (k=10) method.&#x0D; Among the seven algorithms/classifiers, Random Forest had the highest accuracy in schedule performance model and Multilayer Perceptron in quality performance model

High-pressure structural behavior and elastic properties of Ce2In and Ce3In: A combined in situ synchrotron X-ray diffraction and density functional theory study

Across the cerium (Ce)-indium (In) binary phase diagram, there are five line compounds including Ce3In, Ce2In, Ce3In5, CeIn2 and CeIn3. So far, CeIn3 is the only compound whose high-pressure structural behavior and elastic properties have been studied in detail. In this work, we investigated the compression behavior of Ce3In and Ce2In at pressures up to 26.2 GPa using in situ high-pressure angle dispersive synchrotron powder X-ray diffraction (XRD). Our results indicate that Ce3In retains its Cu3Au-type structure in the pressure range investigated, whereas Ce2In undergoes a phase transformation from the Ni2In-type structure to an unidentified phase at ∼4.6 GPa. Unit-cell parameters as a function of pressure were obtained by Rietveld analysis of XRD data. Unit cell volumes were fitted to a second order Birch-Murnaghan equation of state (EOS). The experimentally determined bulk moduli (K0) of the two cerium indides are 50 ± 3 GPa for Ce3In and 39.2 ± 1.6 GPa for Ce2In. The zero-pressure compressibilites of the a-axis (βa,0) for Ce3In is 6.67 (±0.36) × 10–3 GPa–1 and of the a- and c-axis for Ce2In are βa,0 = 8.85 (±0.39) × 10−3 GPa−1 and βc,0 = 9.26 (±0.51) × 10−3 GPa−1, respectively. These results are in general agreement with those calculated by density functional theory (DFT + U) and open-source machine learning algorithms.

AI-based approach for transcribing and classifying unstructured emergency call data: A methodological proposal.

Emergency care-sensitive conditions (ECSCs) require rapid identification and treatment and are responsible for over half of all deaths worldwide. Prehospital emergency care (PEC) can provide rapid treatment and access to definitive care for many ECSCs and can reduce mortality in several different settings. The objective of this study is to propose a method for using artificial intelligence (AI) and machine learning (ML) to transcribe audio, extract, and classify unstructured emergency call data in the Serviço de Atendimento Móvel de Urgência (SAMU) system in southern Brazil. The study used all "1-9-2" calls received in 2019 by the SAMU Novo Norte Emergency Regulation Center (ERC) call center in Maringá, in the Brazilian state of Paraná. The calls were processed through a pipeline using machine learning algorithms, including Automatic Speech Recognition (ASR) models for transcription of audio calls in Portuguese, and a Natural Language Understanding (NLU) classification model. The pipeline was trained and validated using a dataset of labeled calls, which were manually classified by medical students using LabelStudio. The results showed that the AI model was able to accurately transcribe the audio with a Word Error Rate of 42.12% using Wav2Vec 2.0 for ASR transcription of audio calls in Portuguese. Additionally, the NLU classification model had an accuracy of 73.9% in classifying the calls into different categories in a validation subset. The study found that using AI to categorize emergency calls in low- and middle-income countries is largely unexplored, and the applicability of conventional open-source ML models trained on English language datasets is unclear for non-English speaking countries. The study concludes that AI can be used to transcribe audio and extract and classify unstructured emergency call data in an emergency system in southern Brazil as an initial step towards developing a decision-making support tool.

Tree species classification on images from airborne mobile mapping using ML.NET

ABSTRACT Deep learning is a powerful tool for automating the process of recognizing and classifying objects in images. In this study, we used ML.NET, a popular open-source machine learning framework, to develop a model for identifying tree species in images obtained from airborne mobile mapping. These high-resolution images can be used to create detailed maps of the landscape. They can also be analyzed and processed to extract information about visual features, including tree species recognition. The deep learning model was trained using ML.NET to classify two tree species based on the combination of airborne mobile mapping images. Our approach yielded impressive results, with a maximum classification accuracy of 93.9%. This demonstrates the effectiveness of combining imagery sources with deep learning tools in ML.NET for efficient and accurate tree species classification. This study highlights the potential of the ML.NET framework for automating object classification and can provide valuable insights and information for forestry management and conservation efforts. The primary objective of this research was to evaluate the effectiveness of an approach for identifying tree species through a model generated using a combination of ortho and oblique images captured by a mobile mapping system.

Automated maternal behavior during early life in rodents (AMBER) pipeline

Mother-infant interactions during the early postnatal period are critical for infant survival and the scaffolding of infant development. Rodent models are used extensively to understand how these early social experiences influence neurobiology across the lifespan. However, methods for measuring postnatal dam-pup interactions typically involve time-consuming manual scoring, vary widely between research groups, and produce low density data that limits downstream analytical applications. To address these methodological issues, we developed the Automated Maternal Behavior during Early life in Rodents (AMBER) pipeline for quantifying home-cage maternal and mother–pup interactions using open-source machine learning tools. DeepLabCut was used to track key points on rat dams (32 points) and individual pups (9 points per pup) in postnatal day 1–10 video recordings. Pose estimation models reached key point test errors of approximately 4.1–10 mm (14.39 pixels) and 3.44–7.87 mm (11.81 pixels) depending on depth of animal in the frame averaged across all key points for dam and pups respectively. Pose estimation data and human-annotated behavior labels from 38 videos were used with Simple Behavioral Analysis (SimBA) to generate behavior classifiers for dam active nursing, passive nursing, nest attendance, licking and grooming, self-directed grooming, eating, and drinking using random forest algorithms. All classifiers had excellent performance on test frames, with F1 scores above 0.886. Performance on hold-out videos remained high for nest attendance (F1 = 0.990), active nursing (F1 = 0.828), and licking and grooming (F1 = 0.766) but was lower for eating, drinking, and self-directed grooming (F1 = 0.534–0.554). A set of 242 videos was used with AMBER and produced behavior measures in the expected range from postnatal 1–10 home-cage videos. This pipeline is a major advancement in assessing home-cage dam-pup interactions in a way that reduces experimenter burden while increasing reproducibility, reliability, and detail of data for use in developmental studies without the need for special housing systems or proprietary software.

Geometry Optimization: A Comparison of Different Open-Source Geometry Optimizers.

Based on a series of energy minimizations with starting structures obtained from the Baker test set of 30 organic molecules, a comparison is made between various open-source geometry optimization codes that are interfaced with the open-source QUantum Interaction Computational Kernel (QUICK) program for gradient and energy calculations. The findings demonstrate how the choice of the coordinate system influences the optimization process to reach an equilibrium structure. With fewer steps, internal coordinates outperform Cartesian coordinates, while the choice of the initial Hessian and Hessian update method in quasi-Newton approaches made by different optimization algorithms also contributes to the rate of convergence. Furthermore, an available open-source machine learning method based on Gaussian process regression (GPR) was evaluated for energy minimizations over surrogate potential energy surfaces with both Cartesian and internal coordinates with internal coordinates outperforming Cartesian. Overall, geomeTRIC and DL-FIND with their default optimization method as well as with the GPR-based model using Hartree-Fock theory with the 6-31G** basis set needed a comparable number of geometry optimization steps to the approach of Baker using a unit matrix as the initial Hessian to reach the optimized geometry. On the other hand, the Berny and Sella offerings in ASE outperformed the other algorithms. Based on this, we recommend using the file-based approaches, ASE/Berny and ASE/Sella, for large-scale optimization efforts, while if using a single executable is preferable, we now distribute QUICK integrated with DL-FIND.

SpikingJelly: An open-source machine learning infrastructure platform for spike-based intelligence.

Spiking neural networks (SNNs) aim to realize brain-inspired intelligence on neuromorphic chips with high energy efficiency by introducing neural dynamics and spike properties. As the emerging spiking deep learning paradigm attracts increasing interest, traditional programming frameworks cannot meet the demands of the automatic differentiation, parallel computation acceleration, and high integration of processing neuromorphic datasets and deployment. In this work, we present the SpikingJelly framework to address the aforementioned dilemma. We contribute a full-stack toolkit for preprocessing neuromorphic datasets, building deep SNNs, optimizing their parameters, and deploying SNNs on neuromorphic chips. Compared to existing methods, the training of deep SNNs can be accelerated 11×, and the superior extensibility and flexibility of SpikingJelly enable users to accelerate custom models at low costs through multilevel inheritance and semiautomatic code generation. SpikingJelly paves the way for synthesizing truly energy-efficient SNN-based machine intelligence systems, which will enrich the ecology of neuromorphic computing.