Learning Iteration Research Articles

Active learning via adaptive weighted uncertainty sampling applied to additive manufacturing

The acquisition of large amounts of high-dimensional data is becoming prevalent in additive manufacturing, arising from sensor integration in such processes. Although supervised machine learning has gained popularity to predict process quality from such data, annotation can be time-consuming and labor intensive. Active learning is a sub-field within machine learning concerned with maximizing model performance using the least amount of annotated data. The focus of this study is twofold. Firstly, a novel active learning method is introduced, called adaptive weighted uncertainty sampling (AWUS) which balances uncertainty with random sampling using the model change between active learning iterations. This way, AWUS is able to adapt from exploration of the instance space to exploitation of model knowledge throughout the active learning process. Secondly, a novel feature extraction and classification method is proposed for directed-energy-deposition additive manufacturing processes which predicts image and process quality based upon the visibility of the melt-pool using the visual signature of occluding features in the image. AWUS is compared against 6 state-of-the-art query strategies on 28 datasets from the OpenML database and 8 in-situ machine vision recordings of directed-energy-deposition processes using 4 different classifiers. The results show that AWUS outperforms the state-of-the-art across the board of the datasets, classifiers and active learning batch sizes. Furthermore, the application of AWUS reduces the amount of necessary annotations by 20–70% in 90% of our experiments compared to random sampling.

A brief review of simulated Kalman Filter Algorithm – variants and applications

Simulated Kalman Filter (SKF) solves optimization problems by finding the estimate of the optimum solution. As a multi-agent algorithm, every agent in the population acts as a Kalman filter by using a standard Kalman filter framework, which includes a simulated measurement process and a best-so-far solution as a reference. This paper presented an overview of the research progress in SKF from the day it was introduced until the present day, discussing the progress, improvements, modifications, and applications of SKF. The fundamental and standard algorithm were first introduced. Then, the work on the algorithm improvements was surveyed. Finally, the remaining unresolved problems and some directions of SKF research were discussed. We reviewed 57 SKF papers. 16 of them on fundamental improvements, 9 on extension of the algorithm to discrete problems and 25 on their applications. Researchers have worked on ideas to improve exploration capability to prevent premature convergence by trying prediction operators, opposition-based learning, and different iteration strategies. There were also attempts to hybridize SKF with other famous algorithms such as Particle Swarm Optimization (PSO), Gravitational Search Algorithm (GSA), and Sine Cosine Algorithm (SCA) to improve its performance. Lastly, a single-agent variant of SKF and a multi-objective SKF were introduced. SKF algorithms and its variants have been implemented in at least nine areas of applications: drill path optimization, airport gate allocation problem (AGAP), assembly sequence planning (ASP), system identification, feature selection, image template matching, controller tuning, wireless sensor network, and engineering design problem. The literature reviewed solely depended on the keyword search that contained the terms simulated Kalman filter from December 2015 to the present date. This is the first review paper on SKF. It is hoped that this survey would be beneficial for the researchers of this area and attracting interest towards the algorithm.

Automated quantification of epicardial adipose tissue on CCTA via deep-learning detection of the pericardium: clinical implications

Abstract Background Epicardial adipose tissue (EAT) is a visceral fat deposit within the pericardial sac which surrounds the heart myocardium and coronary arteries. EAT volume has been demonstrated to be strongly associated with the development and prognosis of cardiovascular diseases, but its measurement is subjective and challenging in practice. Purpose To develop a deep-learning approach for automated segmentation of EAT from routine CCTA scans, that could assist clinical interpretation of CCTA. Methods A deep-learning method using a 3D Residual-U-Net neural network architecture for 3D volumetric segmentation of CCTA data was created. The network was trained on a diverse sample of 1900 CCTAs, each manually segmented by a single expert, drawn from the UK sites of the Oxford Risk Factors And Non-invasive imaging (ORFAN) Study. Three iterations of feedback learning were used to fine tune the algorithm for the segmentation of the whole heart within the bounds of the pericardium. In each iteration, the machine analysed sets of 100–250 unannotated CCTAs unseen by the machine which were then corrected by experts. EAT volumes were calculated by automated thresholding of adipose tissue (−190HU through −30HU) from within the bound of the pericardial segment (Figure 1). The network was then applied to 817 unseen CCTAs from US sites of the ORFAN Study. These scans were also segmented for ground truth by two experts blind to all other data. Comparisons between machine vs expert total pericardial volume and EAT volume were made using Lin's concordance correlation coefficient (CCC). The algorithm was then applied externally in 1588 CCTAs from the SCOTHEART trial (UK), and the EAT volume was automatically calculated for each case. Cross-sectional associations between standardised EAT volumes and prevalent AF and CAD were performed. Results Within both the internal (UK ORFAN sites) and external (USA ORFAN sites) validation cohorts correlation between human and machine segmented total pericardium and EAT was excellent, with CCC of 0.97 for both volumes (external validation cohort shown in Figure 2A). Utilising SCOTHEART CCTAs with automatically segmented EAT volumes, a multivariable-adjusted logistic regression model accounting for risk factors of age, sex, BMI, hypertension, diabetes mellitus, valvular disease, and previous heart surgery found that EAT volumes were significantly associated with prevalent AF, with odds ratio (OR) per 1 SD increase of EAT volume of 1.20 (95% CI, 1.06 to 1.44; P=0.03). A similar model for prevalent CAD, adjusted for age, sex, BMI, hypertension, non-HDL cholesterol, diabetes mellitus, and coronary artery calcium score resulted in an OR per 1 SD increase of EAT volume of 1.26 (95% CI, 1.10 to 1.45; P=0.001) (Figure 2B). Conclusion Highly accurate, reproducible, and instantaneous EAT volume quantification is possible utilising deep-learning detection of the whole human heart within the pericardial sac. Funding Acknowledgement Type of funding sources: Public Institution(s). Main funding source(s): British Heart FoundationNational Institute for Health Research - Oxford University Hospitals Biomedical Research Centre Figure 1Figure 2

Enhancement of multilayer perceptron model training accuracy through the optimization of hyperparameters: a case study of the quality prediction of injection-molded parts

Injection molding has been broadly used in the mass production of plastic parts and must meet the requirements of efficiency and quality consistency. Machine learning can effectively predict the quality of injection-molded part. However, the performance of machine learning models largely depends on the accuracy of the training. Hyperparameters such as activation functions, momentum, and learning rate are crucial to the accuracy and efficiency of model training. This research aims to analyze the influence of hyperparameters on testing accuracy, explore the corresponding optimal learning rate, and provide the optimal training model for predicting the quality of injection-molded parts. In this study, stochastic gradient descent (SGD) and stochastic gradient descent with momentum (SGDM) are used to optimize the artificial neural network model. Through optimization of these training model hyperparameters, the width testing accuracy of the injection product is improved. The experimental results indicate that in the absence of momentum effects, all five activation functions can achieve more than 90% of the training accuracy with a learning rate of 0.1. Moreover, when optimized with the SGD, the learning rate of the Sigmoid activation function is 0.1, and the testing accuracy reaches 95.8%. Although momentum has the least influence on accuracy, it affects the convergence speed of the Sigmoid function, which reduces the number of required learning iterations (82.4% reduction rate). In summary, optimizing hyperparameter settings can improve the accuracy of model testing and markedly reduce training time.

Pharma 4.0 Continuous mRNA Drug Products Manufacturing.

Continuous mRNA drugs manufacturing is perceived to nurture flow processes featuring quality by design, controlled automation, real time validation, robustness, and reproducibility, pertaining to regulatory harmonization. However, the actual adaptation of the latter remains elusive, hence batch-to-continuous transition would a priori necessitate holistic process understanding. In addition, the cost related to experimental, pilot manufacturing lines development and operations thereof renders such venture prohibitive. Systems-based Pharmaceutics 4.0 digital design enabling tools, i.e., converging mass and energy balance simulations, Monte-Carlo machine learning iterations, and spatial arrangement analysis were recruited herein to overcome the aforementioned barriers. The primary objective of this work is to hierarchically design the related bioprocesses, embedded in scalable devices, compatible with continuous operation. Our secondary objective is to harvest the obtained technological data and conduct resource commitment analysis. We herein demonstrate for first time the feasibility of the continuous, end-to-end production of sterile mRNA formulated into lipid nanocarriers, defining the equipment specifications and the desired operational space. Moreover, we find that the cell lysis modules and the linearization enzymes ascend as the principal resource-intensive model factors, accounting for 40% and 42% of the equipment and raw material, respectively. We calculate MSPD 1.30–1.45 €, demonstrating low margin lifecycle fluctuation.

Optimal game theoretic solution of the pursuit‐evasion intercept problem using on‐policy reinforcement learning

AbstractThis article presents a rigorous formulation for the pursuit‐evasion (PE) game when velocity constraints are imposed on agents of the game or players. The game is formulated as an infinite‐horizon problem using a non‐quadratic functional, then sufficient conditions are derived to prove capture in a finite‐time. A novel tracking Hamilton–Jacobi–Isaacs (HJI) equation associated with the non‐quadratic value function is employed, which is solved for Nash equilibrium velocity policies for each agent with arbitrary nonlinear dynamics. In contrast to the existing remedies for proof of capture in PE game, the proposed method does not assume players are moving with their maximum velocities and considers the velocity constraints a priori. Attaining the optimal actions requires the solution of HJI equations online and in real‐time. We overcome this problem by presenting the on‐policy iteration of integral reinforcement learning (IRL) technique. The persistence of excitation for IRL to work is satisfied inherently until capture occurs, at which time the game ends. Furthermore, a nonlinear backstepping control method is proposed to track desired optimal velocity trajectories for players with generalized Newtonian dynamics. Simulation results are provided to show the validity of the proposed methods.

Multi‐dimensional weighted cross‐attention network in crowded scenes

Human detection in crowded scenes is one of the research components of crowd safety problem analysis, such as emergency warning and security monitoring platforms. Although the existing anchor-free methods have fast inference speed, they are not suitable for object detection in crowded scenes due to the model's inability to predict the well-fined object detection bounding boxes. This work proposes an end-to-end anchor-free network, Multi-dimensional Weighted Cross-Attention Network (MANet), which can perform real-time human detection in crowded scenes. Specifically, the Double-flow Weighted Feature Cascade Module (DW-FCM) is used in the extractor to highlight the contribution of features at different levels. The Triplet Cross Attention Module (TCAM) is used in the detector head to enhance the association dependence of multi-dimension features, further strengthening human boundary features' discrimination ability at a fine-grained level. Moreover, the strategy of Adaptively Opposite Thrust Mapping (AOTM) ground-truth annotation is proposed to achieve bias correction of erroneous mappings and reduce the iterations of useless learning of the network. These strategies effectively alleviate the defect that the existing anchor-free network cannot correctly distinguish and locate the individual human in crowded scenes. Compared with the anchor-based detection method, there is no need to set anchor parameters manually, and the detection speed can satisfy the real-time application. Finally, through extensive comparative experiments on CrowdHuman and WIDER FACE datasets, the results demonstrate that the improved strategy achieves the state-of-the-art result in the anchor-free methods.

Evaluation of Deep Learning Based YOLOv3 Brain Tumor Classification Performance Using Magnetic Resonance Imaging

Magnetic resonance imaging (MRI) provides high-resolution images of soft tissues and is an imaging technique with a high diagnostic value. It can play a role in computer-aided diagnosis through deep learning technologies using digital data. This study aims to investigate the performance of brain tumor classification using YOLOv3 based on deep learning. Deep learning was performed using 253 open MRI images in which the learning evaluation indices were average loss, region 82, and region 94. The detection performance was evaluated using images that were not used for training to verify the brain tumor classification model. The average loss was 0.1107 for 2248 epochs. After 24,079 learning iterations, average IoU, class, .5R, and .75R were 0.89, 0.81, 1.00, and 1.00 for region 82, and 1.00, 1.00, 1.00, and 1.00 for region 94, respectively. Owing to the verification of the brain tumor classification model, it was possible to classify normal brain and brain tumors with an accuracy of 95.00% and 75.36%, respectively. It is believed that the results of this study will be used as basic data for deep learning research and clinical trials using MRI images.

A Study on Pagoda Image Search Using Artificial Intelligence (AI) Technology for Restoration of Cultural Properties

The current cultural assets are being restored depending on the opinions of experts (craftsmen). We intend to introduce digitalized artificial intelligence techniques, excluding the personal opinions of experts on reconstruction of such cultural properties. The first step toward restoring digitized cultural properties is separation. The restoration of cultural properties should be reorganized based on recorded documents, period historical backgrounds and regional characteristics. The cultural properties in the form of photographs or images should be collected by separating the background. In addition, when restoring cultural properties most of them depend a lot on the tendency of the restoring person workers. As a result, it often occurs when there is a problem in the accuracy and reliability of restoration of cultural properties. In this study, we propose a search method for learning stored digital cultural assets using AI technology. Pagoda was selected for restoration of Cultural Properties. Pagoda data collection was collected through the Internet and various historical records. The pagoda data was classified by period and region, and grouped into similar buildings. The collected data was learned by applying the well-known CNN algorithm for artificial intelligence learning. The pagoda search used Yolo Marker to mark the tower shape. The tower was used a total of about 100-10,000 pagoda data. In conclusion, it was confirmed that the probability of searching for a tower differs according to the number of pagoda pictures and the number of learning iterations. Finally, it was confirmed that the number of 500 towers and the epochs in training of 8000 times were good. If the test result exceeds 8,000 times, it becomes overfitting. All so, I found a phenomenon that the recognition rate drops when the enemy repeatedly learns more than 8,000 times. As a result of this study, it is believed that it will be helpful in data gathering to increase the accuracy of tower restoration.

Active cluster annotation for wafer map pattern classification in semiconductor manufacturing

In semiconductor manufacturing, wafer map pattern classification (WMPC) is important for ensuring good manufacturing quality because the defect type on a wafer map provides important information for determining defect causes. To construct a high-performance WMPC model, a large amount of labeled training data is required, which entails a high annotation cost for engineers. To reduce the annotation cost, an active learning framework has been investigated, in which annotation is conducted at the individual wafer map level. If wafer maps can be grouped into clusters based on the similarity of defect patterns, then annotation can be performed at the cluster level rather than at the wafer map level, thereby affording cost effectiveness. Based on the cluster-level annotation, we propose an active cluster annotation to obtain a high-performance WMPC model with reduced annotation cost. For a dataset annotated only for a small subset of wafer maps, clustering is first conducted for unlabeled wafer maps. In an active learning iteration, a convolutional neural network (CNN) is constructed with labeled wafer maps. Subsequently, the cluster-level classification uncertainties of the CNN for unlabeled wafer maps are calculated. With the uncertainties, the query clusters are selected and annotated by an engineer at the cluster level. The performance of the CNN is improved cost-effectively by repeating these iterations. We demonstrate the effectiveness of the proposed method through experiments on real-world data from a semiconductor manufacturer. In addition, we show that cluster-level annotation is a robust annotation method that can yield consistent labels.

Learning Time Reduction Using Warm-Start Methods for a Reinforcement Learning-Based Supervisory Control in Hybrid Electric Vehicle Applications

Reinforcement Learning (RL) is widely utilized in the field of robotics, and as such, it is gradually being implemented in the Hybrid Electric Vehicle (HEV) supervisory control. Even though RL exhibits excellent performance in terms of fuel consumption minimization in simulation, the large learning iteration number needs a long learning time, making it hardly applicable in real-world vehicles. In addition, the fuel consumption of initial learning phases is much worse than baseline controls. This study aims to reduce the learning iterations of Q-learning in HEV application and improve fuel consumption in initial learning phases utilizing warm start methods. Different from previous studies, which initiated Q-learning with zero or random Q values, this study initiates the Q-learning with different supervisory controls (i.e., Equivalent Consumption Minimization Strategy control and heuristic control), and detailed analysis is given. The results show that the proposed warm start Q-learning requires 68.8% fewer iterations than cold start Q-learning. The trained Q-learning is validated in two different driving cycles, and the results show 10-16% MPG improvement when compared to Equivalent Consumption Minimization Strategy control. Furthermore, real-time feasibility is analyzed, and the guidance of vehicle implementation is provided. The results of this study can be used to facilitate the deployment of RL in vehicle supervisory control applications.

Determining 1D fast-ion velocity distribution functions from ion cyclotron emission data using deep neural networks

The relationship between simulated ion cyclotron emission (ICE) signals s and the corresponding 1D velocity distribution function fv⊥ of the fast ions triggering the ICE is modeled using a two-layer deep neural network. The network architecture (number of layers and number of computational nodes in each layer) and hyperparameters (learning rate and number of learning iterations) are fine-tuned using a bottom-up approach based on cross-validation. Thus, the optimal mapping gs;θ of the neural network in terms of the number of nodes, the number of layers, and the values of the hyperparameters, where θ is the learned model parameters, is determined by comparing many different configurations of the network on the same training and test set and choosing the best one based on its average test error. The training and test sets are generated by computing random ICE velocity distribution functions f and their corresponding ICE signals s by modeling the relationship as the linear matrix equation Wf = s. The simulated ICE signals are modeled as edge ICE signals at LHD. The network predictions for f based on ICE signals s are on many simulated ICE signal examples closer to the true velocity distribution function than that obtained by 0th-order Tikhonov regularization, although there might be qualitative differences in which features one technique is better at predicting than the other. Additionally, the network computations are much faster. Adapted versions of the network can be applied to future experimental ICE data to infer fast-ion velocity distribution functions.

Awesome back-propagation machine learning paradigm

For a better future in machine learning (ML), it is necessary to modify our current concepts to get the fastest ML. Many designers had attempted to find the optimal learning rates in their applications through many algorithms over the past decades, but they have not yet achieved their target of highest speed of back-propagation (BP). This research proposes a novel BP rule called the Instant Learning Ratios-Machine Learning (ILR-ML) or (ILRML). Unlike the traditional BP algorithms, the ILR-ML offers its learning without the concepts of the learning rate(s). The ILR-ML has a new concept called the "Learning Ratio" and indicated by a sign (Δl). The ILR-ML performs the full BP algorithm with 100% accuracy per each learning iteration. The ILR-ML is more suitable for the online machine learning.

Cost-sensitive semi-supervised deep learning to assess driving risk by application of naturalistic vehicle trajectories

Most traffic accidents are caused by driver-related factors such as poor perception, aggressive decision-making, or improper maneuvering. Therefore, it is critical to evaluate and predict driving risks to provide drivers with timely feedback. However, risk assessment involves challenges related to a lack of labeled driving data and the presence of data imbalance in the description of different driving risk levels. To address these challenges, a cost-sensitive semi-supervised deep learning method is developed to obtain driving risk scores based on naturalistic vehicle trajectories. A convolutional neural network and a long short-term memory encoder/decoder network are embedded into a semi-supervised framework that uses only a small labeled dataset to label the remaining unlabeled data and produce a trained network model. As fixed weights cannot adapt to changes in the degree of class imbalance that occur over progressive semi-supervised learning iterations, an adaptive over-balanced cross-entropy loss function is developed to adaptively maintain an over-balanced state for the high-risk class to achieve cost-sensitive learning. The experimental results indicate that the accuracy of the proposed method in determining the current and future 2 s risk scores is 96.63% and 92.06%, respectively, thereby constituting the best comprehensive performance among existing machine learning methods. Moreover, the method is verified using a spatio-temporal diagram of driving risk-trajectory and a current–future risk score diagram. The findings demonstrate that the proposed method can be used to assess driving risks in a reliable and robust manner.

Analysis of the Resilient Method in Training and Accuracy in the Backpropagation Method

Artificial Neural Network (ANN) is one of the clusters of computer science that leads to artificial intelligence, there are several methods in ANN, one of which is the backpropagation method. This method is used in the prediction process. In some cases the backpropagation method can help in problems solving, especially predictions. However, the backpropagation method has weaknesses. The results of the backpropagation method are very influenced by the determination of the parameters so that the convergence becomes very slow. So needed an optimization method to optimize the performance of the bakpropagation method. The resilient backpropgation method is one solution, this method can change the weight and bias of the network with a direct adaptation process of weighting based on local gradient information from learning iterations so that it can provide optimal results. The data used is the Higher Education Gross Enrollment Rate in Indonesia from 2015-2020 by province. The results were obtained from several data testing with architectural experiments 3-5-1, 3-20-1, 3-37-1, 3-19-1, 3-26-4 and 3-4-1 from backpropagation and resilient testing, shows that the data training process can be optimized significantly, but the accuracy is not evenly optimal

Patch-Based Change Detection Method for SAR Images with Label Updating Strategy

Convolutional neural networks (CNNs) have been widely used in change detection of synthetic aperture radar (SAR) images and have been proven to have better precision than traditional methods. A two-stage patch-based deep learning method with a label updating strategy is proposed in this paper. The initial label and mask are generated at the pre-classification stage. Then a two-stage updating strategy is applied to gradually recover changed areas. At the first stage, diversity of training data is gradually restored. The output of the designed CNN network is further processed to generate a new label and a new mask for the following learning iteration. As the diversity of data is ensured after the first stage, pixels within uncertain areas can be easily classified at the second stage. Experiment results on several representative datasets show the effectiveness of our proposed method compared with several existing competitive methods.

Blended Learning and the Videoconferencing Modality for the Teaching and Learning Doctoral Candidate

This paper simply documents the changes to a foundations course for teaching and learning doctoral candidates at a Midwest university, required due to the recent pandemic conditions and subsequent healthcare precautions. The main change to this already blended modality course, Critical Policy Analysis, involved modifying three face-to-face (F2F), in person sessions, essentially transforming them into three videoconferencing sessions using the ZOOM application. My reflection provides what can only be considered to be a discussion of very basic findings, but amongst those findings are that digital tools often appear to ‘work’ very seamlessly and hence exceed our collective expectations. The breakout sessions structured into the course, for example, provided a real-time experience comparable in most ways to small group discussions held previously when the course was offered in a physical classroom. However, some of the students struggled with the online portion of the class (course modules, with electronic reserve readings, discussion posts, and crafting a white paper/report, all online) and felt at times less than proficient working online. A few indicated preference for classes in person.Also included in this paper are some of my observations as a long-time online instructor and researcher in the educational communications technology field. I argue that this latest iteration of blended learning is here to stay, and that this change will pose both promise and peril for education in the years ahead.

Deep active learning for classifying cancer pathology reports

BackgroundAutomated text classification has many important applications in the clinical setting; however, obtaining labelled data for training machine learning and deep learning models is often difficult and expensive. Active learning techniques may mitigate this challenge by reducing the amount of labelled data required to effectively train a model. In this study, we analyze the effectiveness of 11 active learning algorithms on classifying subsite and histology from cancer pathology reports using a Convolutional Neural Network as the text classification model.ResultsWe compare the performance of each active learning strategy using two differently sized datasets and two different classification tasks. Our results show that on all tasks and dataset sizes, all active learning strategies except diversity-sampling strategies outperformed random sampling, i.e., no active learning. On our large dataset (15K initial labelled samples, adding 15K additional labelled samples each iteration of active learning), there was no clear winner between the different active learning strategies. On our small dataset (1K initial labelled samples, adding 1K additional labelled samples each iteration of active learning), marginal and ratio uncertainty sampling performed better than all other active learning techniques. We found that compared to random sampling, active learning strongly helps performance on rare classes by focusing on underrepresented classes.ConclusionsActive learning can save annotation cost by helping human annotators efficiently and intelligently select which samples to label. Our results show that a dataset constructed using effective active learning techniques requires less than half the amount of labelled data to achieve the same performance as a dataset constructed using random sampling.

Premature Ventricular Contractions’ Detection Based on Active Learning

Premature ventricular contractions (PVCs) are one of the most common cardiovascular diseases with high risk to a large population of patients. It has been shown that supervised learning algorithms can detect PVCs from beat-level ECG data. However, a huge human effort is needed in order to achieve an accurate detection rate. A convolutional autoencoder was trained in this work in an unsupervised fashion to extract features automatically with zero prior specialized knowledge. Random forest was adopted as a supervised algorithm trained on the features generated by the autoencoder. Various active learning selection strategies, uncertainty-based and diversity-based, were studied on top of the random forest. In each iteration of active learning, the training data are updated with newly selected samples and fed into the classifier. The performance on an independent validation set is recorded in each iteration. As a result, among the different uncertainty sampling strategies, the least confidence score shows a better F1 score of 0.85 than other methods. In between the two diversity-based strategies, the representative clustering sample had the best F1 score than the k-center-greedy algorithm. By comparing the performance of different active learning methods trained on half of the original data size with the same classifier trained on the full set, the F1 score of least confidence is still better than the full set. This study demonstrates that active learning could help reduce human annotation effort by achieving the same level of performance as the classifier trained on the fully annotated training data.

Rheology-Based Classification of Foods for the Elderly by Machine Learning Analysis

A new research framework for the rheological measurements of foods for the elderly was proposed by combining experiments with machine learning. Universal design food (UDF), the conventional rheological test for foods for the elderly, was compared with three different rheological methods in terms of stress, clearly showing a great linear correlation (R2 = 0.9885) with the puncture test. A binary logistic classification with the tensorflow library was successfully applied to predict the elderly’s foods based on the rheological stress values from the UDF and puncture tests. The gradient descent algorithm demonstrated that the cost functions became minimized, and the model parameters were optimally estimated with an increasing number of machine learning iterations. From the testing dataset, the predictive model with a threshold value of 0.7 successfully classified the food samples into two groups (belong to the elderly’s foods or not) with an accuracy of 98%. The research framework proposed in this study can be applied to a wide variety of classification and estimation-related studies in the field of food science.