Search

Machine learning at the extreme edge has enabled a plethora of intelligent, time-critical, and remote applications. However, deploying interpretable artificial intelligence systems that can perform high-level symbolic reasoning and satisfy the underlying system rules and physics within the tight platform resource constraints is challenging. In this paper, we introduce TinyNS, the first platform-aware neurosymbolic architecture search framework for joint optimization of symbolic and neural operators. TinyNS provides recipes and parsers to automatically write microcontroller code for five types of neurosymbolic models, combining the context awareness and integrity of symbolic techniques with the robustness and performance of machine learning models. TinyNS uses a fast, gradient-free, black-box Bayesian optimizer over discontinuous, conditional, numeric, and categorical search spaces to find the best synergy of symbolic code and neural networks within the hardware resource budget. To guarantee deployability, TinyNS talks to the target hardware during the optimization process. We showcase the utility of TinyNS by deploying microcontroller-class neurosymbolic models through several case studies. In all use cases, TinyNS outperforms purely neural or purely symbolic approaches while guaranteeing execution on real hardware.

Automated Machine Learning (AutoML) tools are revolutionizing the field of machine learning by significantly reducing the need for deep computer science expertise. Designed to make ML more accessible, they enable users to build high-performing models without extensive technical knowledge. This study delves into these tools in the context of time series analysis, which is essential for forecasting future trends from historical data. We evaluate three prominent AutoML tools—AutoGluon, Auto-Sklearn, and PyCaret—across various metrics, employing diverse datasets that include Bitcoin and COVID-19 data. The results reveal that the performance of each tool is highly dependent on the specific dataset and its ability to manage the complexities of time series data. This thorough investigation not only demonstrates the strengths and limitations of each AutoML tool but also highlights the criticality of dataset-specific considerations in time series analysis. Offering valuable insights for both practitioners and researchers, this study emphasizes the ongoing need for research and development in this specialized area. It aims to serve as a reference for organizations dealing with time series datasets and a guiding framework for future academic research in enhancing the application of AutoML tools for time series forecasting and analysis.

Addressing the need for multiple health indicators is critical to improving prediction accuracy and reducing the limitation of reliance on a single health indicator. This paper presents an AutoML model for accurately forecasting the life span of lithium-ion batteries (LIBs) in electric vehicles. Unlike previous studies focusing solely on constant current (CC) and constant voltage (CV) charging, our AutoML approach includes ICA and direct analysis to generate eight novel multi-dimensional health indicators. These novel comprehensive indicators characterizing battery aging, including the position values of CC charging (IC_CC_P), the peak of CV charging (IC_CV_ H), and the peak/position of CC discharging (IC_D_H, IC_D_ P), are chosen based on correlation analysis and combined with four health indicators from direct analysis, i.e., capacity, cycle, cluster, and checkpoint, to improve prediction accuracy. K-means clustering is employed to group similar data points. At the same time, the ruptures library facilitates the implementation of a novel algorithm for change point detection, allowing for the identification of checkpoints within each battery dataset. Experimental validation using NASA's AMES LIB cycle life datasets demonstrates a significant performance improvement from existing recent methods, with 58.93 % lower mean absolute error (MAE) and 52.66 % lower root-mean-square error (RMSE) compared to recent AutoML-based methods. This shows a significant improvement in forecasting the RUL of electric vehicle LIBs using the proposed AutoML-based approach with novel multi-dimensional health indicators. [ABSTRACT FROM AUTHOR]

Digitalization is continuing facilitating our daily lives. The world is interconnected as never before, bringing close people, businesses, or other organizations. However, hackers are also coming close. New business and operational models require the collection and processing of massive amounts of data in real-time, involving utilization of complex information systems, large supply-chains, personal devices, etc. These impose several advantages for adversaries on the one hand (e.g., poorly protected or monitored elements, slow fashion of security updates/upgrades in components that gain little attention, etc.), and many difficulties for defenders on the other hand (e.g., administrate large and complex systems with high dynamicity) in this cyber-security interplay. Impactful attacks on ICT systems, critical infrastructures, and supply networks, as well as cyber-warfare are deriving the necessity for more effective defensives. This paper presents a swarm-intelligence solution for incident handling and response. Cyber Threat Intelligence (CTI) is continuously integrated in the system (i.e., MISP, CVEs, STIX, etc.), and Artificial Intelligence (AI)/Machine Learning (ML) are incorporated in the risk assessment and event evaluation processes. Several incident handling and response sub-procedures are automated, improving effectiveness and decreasing response time. Information concerning identified malicious activity is circulated back to the community (i.e., via the MISP information sharing platform) in an open loop. The proposal is applied in the supply-chain of healthcare organizations in Europe (considering also EU data protection regulations). Nevertheless, it is a generic solution that can be applied in any domain. [ABSTRACT FROM AUTHOR]

This research undertakes an exhaustive exploration of the Auto Machine Learning-based approach (AutoML) to ascertain the source printer for microscopic printed image forensics. The primary focus is on enhancing precision and efficacy in a scenario with limited sample data, achieved by integrating traditional image processing methods with diverse AutoML representatives. The investigation involves a meticulous evaluation and comparative analysis of classical Machine Learning (ML) models and AutoML models in microscopic printed image forensics. Through extensive experiments, we highlight the inherent limitations of traditional models and underscore the compelling advantages offered by AutoML. Notably, we recognize AutoML's unique capability to discern varying degrees of uncertainty within printed patterns. The outcomes emphasize the prowess of Auto Machine Learning in improving the accuracy of printed image forensics, making it particularly promising for future research, especially given the tabular nature of the data and the imperative for a lightweight model suitable for embedded systems. [ABSTRACT FROM AUTHOR]

Knowledge of the factors influencing nutrient-limited subtropical maize yield and subsequent prediction is crucial for effective nutrient management, maximizing profitability, ensuring food security, and promoting environmental sustainability. We analyzed data from nutrient omission plot trials (NOPTs) conducted in 324 farmers' fields across ten agroecological zones (AEZs) in the Eastern Indo-Gangetic Plains (EIGP) of Bangladesh to explain maize yield variability and identify variables controlling nutrient-limited yields. An additive main effect and multiplicative interaction (AMMI) model was used to explain maize yield variability with nutrient addition. Interpretable machine learning (ML) algorithms in automatic machine learning (AutoML) frameworks were subsequently used to predict attainable yield relative nutrient-limited yield (RY) and to rank variables that control RY. The stack-ensemble model was identified as the best-performing model for predicting RYs of N, P, and Zn. In contrast, deep learning outperformed all base learners for predicting RYK. The best model's square errors (RMSEs) were 0.122, 0.105, 0.123, and 0.104 for RYN, RYP, RYK, and RYZn, respectively. The permutation-based feature importance technique identified soil pH as the most critical variable controlling RYN and RYP. The RYK showed lower in the eastern longitudinal direction. Soil N and Zn were associated with RYZn. The predicted median RY of N, P, K, and Zn, representing average soil fertility, was 0.51, 0.84, 0.87, and 0.97, accounting for 44, 54, 54, and 48% upland dry season crop area of Bangladesh, respectively. Efforts are needed to update databases cataloging variability in land type inundation classes, soil characteristics, and INS and combine them with farmers' crop management information to develop more precise nutrient guidelines for maize in the EIGP.

This research undertakes an exhaustive exploration of the Auto Machine Learning-based approach (AutoML) to ascertain the source printer for microscopic printed image forensics. The primary focus is on enhancing precision and efficacy in a scenario with limited sample data, achieved by integrating traditional image processing methods with diverse AutoML representatives. The investigation involves a meticulous evaluation and comparative analysis of classical Machine Learning (ML) models and AutoML models in microscopic printed image forensics. Through extensive experiments, we highlight the inherent limitations of traditional models and underscore the compelling advantages offered by AutoML. Notably, we recognize AutoML’s unique capability to discern varying degrees of uncertainty within printed patterns. The outcomes emphasize the prowess of Auto Machine Learning in improving the accuracy of printed image forensics, making it particularly promising for future research, especially given the tabular nature of the data and the imperative for a lightweight model suitable for embedded systems.

Transformer requires a fixed number of layers and heads which makes them inflexible to the complexity of individual samples and expensive in training and inference. To address this, we propose a sample-based Dynamic Hierarchical Transformer (DHT) model whose layers and heads can be dynamically configured with single data samples via solving contextual bandit problems. To determine the number of layers and heads, we use the Uniform Confidence Bound algorithm while we deploy combinatorial Thompson Sampling in order to select specific head combinations given their number. Different from previous work that focuses on compressing trained networks for inference only, DHT is not only advantageous for adaptively optimizing the underlying network architecture during training but also has a flexible network for efficient inference. To the best of our knowledge, this is the first comprehensive data-driven dynamic transformer without any additional auxiliary neural networks that implement the dynamic system. According to the experiment results, we achieve up to 74% computational savings for both training and inference with a minimal loss of accuracy.

Identifying patients with left ventricular ejection fraction (EF), either reduced [EF < 40% (rEF)], mid-range [EF 40–50% (mEF)], or preserved [EF > 50% (pEF)], is considered of primary clinical importance. An end-to-end video classification using AutoML in Google Vertex AI was applied to echocardiographic recordings. Datasets balanced by majority undersampling, each corresponding to one out of three possible classifications, were obtained from the Standford EchoNet-Dynamic repository. A train–test split of 75/25 was applied. A binary video classification of rEF vs. not rEF demonstrated good performance (test dataset: ROC AUC score 0.939, accuracy 0.863, sensitivity 0.894, specificity 0.831, positive predicting value 0.842). A second binary classification of not pEF vs. pEF was slightly less performing (test dataset: ROC AUC score 0.917, accuracy 0.829, sensitivity 0.761, specificity 0.891, positive predicting value 0.888). A ternary classification was also explored, and lower performance was observed, mainly for the mEF class. A non-AutoML PyTorch implementation in open access confirmed the feasibility of our approach. With this proof of concept, end-to-end video classification based on transfer learning to categorize EF merits consideration for further evaluation in prospective clinical studies. [ABSTRACT FROM AUTHOR]

Denial of Service (DoS) attacks target the availability part of the CIA triad (Confidentiality, Availability, and Integrity). A special category of these attacks is the Distributed DoS (DDoS) attack, where the attacker uses a network of compromised systems called a botnet to flood a target server with requests and refuses to serve legitimate users. DDoS attacks can cost an organization millions of dollars in terms of lost revenue, remediation costs, and damage to brand reputation. Hence, all organizations need speedy real-time detection of DDoS attacks. This work presents a DDoS detection and classification framework using the flow-based approach for feature engineering and the AutoML technique. Our detection system is trained on the latest DDoS datasets – CIC-DDoS 2019 and CIC-IDS 2017, which contain various categories of DDoS attacks. We use various tools to perform adversarial attacks on our trained model. We retrain our models using adversarially crafted network packet captures and then test our models for robustness against practical adversarial attacks that an attacker might use to evade detection. Finally, we deploy our model in real-time using a GUI-based tool. Our model achieves a validation accuracy of 99.9% and a low false positive rate of 0.05%. [ABSTRACT FROM AUTHOR]

The enhancement of fabric quality prediction in the textile manufacturing sector is achieved by utilizing information derived from sensors within the Internet of Things (IoT) and Enterprise Resource Planning (ERP) systems linked to sensors embedded in textile machinery. The integration of Industry 4.0 concepts is instrumental in harnessing IoT sensor data, which, in turn, leads to improvements in productivity and reduced lead times in textile manufacturing processes. This study addresses the issue of imbalanced data pertaining to fabric quality within the textile manufacturing industry. It encompasses an evaluation of seven open-source automated machine learning (AutoML) technologies, namely FLAML (Fast Lightweight AutoML), AutoViML (Automatically Build Variant Interpretable ML models), EvalML (Evaluation Machine Learning), AutoGluon, H2OAutoML, PyCaret, and TPOT (Tree-based Pipeline Optimization Tool). The most suitable solutions are chosen for certain circumstances by employing an innovative approach that finds a compromise among computational efficiency and forecast accuracy. The results reveal that EvalML emerges as the top-performing AutoML model for a predetermined objective function, particularly excelling in terms of mean absolute error (MAE). On the other hand, even with longer inference periods, AutoGluon performs better than other methods in measures like mean absolute percentage error (MAPE), root mean squared error (RMSE), and r-squared. Additionally, the study explores the feature importance rankings provided by each AutoML model, shedding light on the attributes that significantly influence predictive outcomes. Notably, sin/cos encoding is found to be particularly effective in characterizing categorical variables with a large number of unique values. This study includes useful information about the application of AutoML in the textile industry and provides a roadmap for employing Industry 4.0 technologies to enhance fabric quality prediction. The research highlights the importance of striking a balance between predictive accuracy and computational efficiency, emphasizes the significance of feature importance for model interpretability, and lays the groundwork for future investigations in this field. [ABSTRACT FROM AUTHOR]

Numerous real-world data contain missing values, while in contrast, most Machine Learning (ML) algorithms assume complete datasets. For this reason, several imputation algorithms have been proposed to predict and fill in the missing values. Given the advances in predictive modeling algorithms tuned in an Automated Machine Learning context (AutoML) setting, a question that naturally arises is to what extent sophisticated imputation algorithms (e.g., Neural Network based) are really needed, or we can obtain a descent performance using simple methods like Mean/Mode (MM). In this article, we experimentally compare six state-of-the-art representatives of different imputation algorithmic families from an AutoML predictive modeling perspective, including a feature selection step and combined algorithm and hyper-parameter selection. We used a commercial AutoML tool for our experiments, in which we included the selected imputation methods. Experiments ran on 25 binary classification real-world incomplete datasets with missing values and 10 binary classification complete datasets in which synthetic missing values are introduced according to different missingness mechanisms, at varying missing frequencies. The main conclusion drawn from our experiments is that the best method on average is the Denoise AutoEncoder on real-world datasets and the MissForest in simulated datasets, followed closely by MM. In addition, binary indicator variables encoding missingness patterns actually improve predictive performance, on average. Last, although there are cases where Neural-Network-based imputation significantly improves predictive performance, this comes at a great computational cost and requires measuring all feature values to impute new samples. [ABSTRACT FROM AUTHOR]

The impact of artificial intelligence in our society is important due to the innovation of processes through data science to know the academic and sociodemographic factors that contribute to late payments in university students, to identify them and make timely decisions for implementing prevention and correction programs, avoiding student dropout due to this economic problem, and ensuring success in their education in a meaningful and focused way. In this sense, the research aims to compare the performance metrics of classification models for late payments in students of a private university by using AutoML algorithms from various existing platforms and solutions such as AutoKeras, AutoGluon, HyperOPT, MLJar, and H2O in a data set consisting of 8,495 records and the application of data balancing techniques. From the implementation and execution of various algorithms, similar metrics have been obtained based on the parameters and optimization functions used automatically by each tool, providing better performance to the H2O platform through the Stacked Ensemble algorithm with metrics accuracy = 0.778. F1 = 0.870, recall = 0.904 and precision = 0.839. The research can be extended to other contexts or areas of knowledge due to the growing interest in automated machine learning, providing researchers with a valuable tool in data science without the need for deep knowledge. [ABSTRACT FROM AUTHOR]

Optimizing a machine learning pipeline for a task at hand requires careful configuration of various hyperparameters, typically supported by an AutoML system that optimizes the hyperparameters for the given training dataset. Yet, depending on the AutoML system's own second-order meta-configuration, the performance of the AutoML process can vary significantly. Current AutoML systems cannot automatically adapt their own configuration to a specific use case. Further, they cannot compile user-defined application constraints on the effectiveness and efficiency of the pipeline and its generation. In this paper, we propose Caml, which uses meta-learning to automatically adapt its own AutoML parameters, such as the search strategy, the validation strategy, and the search space, for a task at hand. The dynamic AutoML strategy of Caml takes user-defined constraints into account and obtains constraint-satisfying pipelines with high predictive performance. [ABSTRACT FROM AUTHOR]

Certain impacts of climate change could potentially be linked to alterations in rainfall patterns, including shifts in rainfall intensity or drought occurrences. Hence, predicting droughts can provide valuable assistance in mitigating the detrimental consequences associated with water scarcity, particularly in agricultural areas or densely populated urban regions. Employing predictive models to calculate drought indices can be a useful method for the effective characterization of drought conditions. This study applied an Auto-Machine-Learning approach to deploy Artificial Neural Network models, aiming to predict the Standardized Precipitation Index in four regions of Zacatecas, Mexico. Climatological time-series data spanning from 1979 to 2020 were utilized as predictive variables. The best models were found using performance metrics that yielded a Mean Squared Error, Mean Absolute Error, and Coefficient of Determination ranging from 0.0296 to 0.0388, 0.1214 to 0.1355, and 0.9342 to 0.9584, respectively, for the regions under study. As a result, the Auto-Machine-Learning approach successfully developed and tested Artificial Neural Network models that exhibited notable predictive capabilities when estimating the monthly Standardized Precipitation Index within the study region. [ABSTRACT FROM AUTHOR]

This paper (i) explores the internal structure of two quantum mechanics datasets (QM7b, QM9), composed of several thousands of organic molecules and described in terms of electronic properties, and (ii) further explores an inverse design approach to molecular design consisting of using machine learning methods to approximate the atomic composition of molecules, using QM9 data. Understanding the structure and characteristics of this kind of data is important when predicting the atomic composition from physical‐chemical properties in inverse molecular designs. Intrinsic dimension analysis, clustering, and outlier detection methods were used in the study. They revealed that for both datasets the intrinsic dimensionality is several times smaller than the descriptive dimensions. The QM7b data is composed of well‐defined clusters related to atomic composition. The QM9 data consists of an outer region predominantly composed of outliers, and an inner, core region that concentrates clustered inliner objects. A significant relationship exists between the number of atoms in the molecule and its outlier/inliner nature. The spatial structure exhibits a relationship with molecular weight. Despite the structural differences between the two datasets, the predictability of variables of interest for inverse molecular design is high. This is exemplified by models estimating the number of atoms of the molecule from both the original properties and from lower dimensional embedding spaces. In the generative approach the input is given by a set of desired properties of the molecule and the output is an approximation of the atomic composition in terms of its constituent chemical elements. This could serve as the starting region for further search in the huge space determined by the set of possible chemical compounds. The quantum mechanic's dataset QM9 is used in the study, composed of 133,885 small organic molecules and 19 electronic properties. Different multi‐target regression approaches were considered for predicting the atomic composition from the properties, including feature engineering techniques in an auto‐machine learning framework. High‐quality models were found that predict the atomic composition of the molecules from their electronic properties, as well as from a subset of only 52.6% size. Feature selection worked better than feature generation. The results validate the generative approach to inverse molecular design. [ABSTRACT FROM AUTHOR]

Sentiment analysis from students' interactions with learning environments is a topic of interest for researchers in the field of education because it can make important contributions to improving the quality of instructional processes through recommendation systems integrated into learning applications, or by improving the quality of courses, by grouping students according to their common interests and providing feedback on school progress. There are two approaches to sentiment analysis: one lexicon-based and another that uses machine learning. In this study, we present a sentiment analysis from two own data sets that represent students' opinions about school. Our goal is to create a model that helps us to automatically label students' opinions, assigning sentiment scores between 0 and 4 (0 for an extremely negative opinion). To train and evaluate the performance of the model, we used opinions collected from 1443 Romanian high school students. The novelty that we propose is the manual labeling system. Our current research which uses a machine learning approach to classify students' opinions obtains an accuracy of 86.507%. [ABSTRACT FROM AUTHOR]

Hyperparameter optimization (HPO) methods alleviate the significant effort required to obtain hyperparameters that perform optimally on visual learning problems. Existing methods are computationally inefficient because they are task agnostic (i.e., they do not adapt to a given task). We present HyperSTAR (System for Task Aware Hyperparameter Recommendation), a task-aware HPO algorithm that improves HPO efficiency for a target dataset by using prior knowledge from previous hyperparameter searches to recommend effective hyperparameters conditioned on the target dataset. HyperSTAR ranks and recommends hyperparameters by predicting their performance on the target dataset. To do so, it learns a joint dataset-hyperparameter space in an end-to-end manner that enables its performance predictor to use previously found effective hyperparameters for other similar tasks. The hyperparameter recommendations of HyperSTAR combined with existing HPO techniques lead to a task-aware HPO system that reduces the time to find the optimal hyperparameters for the target learning problem. Our experiments on image classification, object detection, and model pruning validate that HyperSTAR reduces the evaluation of different hyperparameter configurations by about 50 % compared to existing methods and, when combined with Hyperband, uses only 25 % of the budget required by the vanilla Hyperband and Bayesian Optimized Hyperband to achieve the best performance. [ABSTRACT FROM AUTHOR]

Evaluation of slope failures, which cause significant loss of life and property comparable to natural disasters such as earthquakes, floods and hurricanes, is one of the main areas of interest in geotechnical engineering. Although traditional and modern methods have been developed for slope stability analysis, the importance given to computer-based approaches has increased in recent years. In this study, we investigated the effectiveness of advanced machine learning (ML) algorithms in classification-based slope stability assessment. In this context, examining the impact of input parameters, such as slope height, slope angle, unit volume weight, internal friction angle of the soil, cohesion of the slope material, and water pressure ratio on the slope stability potential and a comparative analysis was performed on the ML algorithms. On the other hand, automated machine learning (AutoML) approaches were used to make rapid and comprehensive comparisons of ensemble, boosting, bagging and traditional ML algorithms to simplifying application development. The weighted ensemble learning algorithm provided by the AutoGluon package outperformed other algorithms in both testing and training accuracy, achieving an impressive rate of 97.5%, according to the obtained results. All algorithms included in the study performed well, with NeuralNetTorch and CatBoost securing the second position with an accuracy rate of 95%. Furthermore, when evaluating the importance of features using the best algorithm, it can be seen that unit volume weight and internal friction angle of soil had the highest weights, 0.225 and 0.200, respectively, indicating their importance in classifying slope stability. In conclusion, our research significantly advanced slope stability assessment, achieving one of the highest accuracy of 0.975 among various classification-based studies. [ABSTRACT FROM AUTHOR]

Companies must make forecasts for the future to take necessary precautions, as well as to guard or expand their position and remain competitive. The development of data technologies has made it easier to reach meaningful data. Analyzing these data with methods such as artificial intelligence, machine learning, and deep learning makes it possible to obtain highly accurate results in future forecasts. However, the presence of numerous methods in the literature poses several challenges for researchers, including selecting the most suitable method and determining the appropriate techniques for model and hyper-parameter selection. Moreover, comparing different values in the model and making hyper-parameter selections can be tedious and time-consuming. Therefore, this study aims to use the AutoML method, which is an advanced version of machine learning. AutoML automates machine learning models, allowing the use and development of machine learning algorithms without requiring expertise in this field. The study carried out forecasts using 6 different AutoML libraries on univariate time series datasets, and forecasting successes were compared over various performance metrics. According to the results obtained on the data set used, it was observed that the Auto_ARIMA library had the highest forecasting success rate among the selected libraries. [ABSTRACT FROM AUTHOR]

Purpose: GraphQL is a new Open API specification that allows clients to send queries and obtain data flexibly according to their needs. However, a high-complexity GraphQL query may lead to an excessive data volume of the query result, which causes problems such as resource overload of the API server. Therefore, this paper aims to address this issue by predicting the response data volume of a GraphQL query statement. Design/methodology/approach: This paper proposes a GraphQL response data volume prediction approach based on Code2Vec and AutoML. First, a GraphQL query statement is transformed into a path collection of an abstract syntax tree based on the idea of Code2Vec, and then the query is aggregated into a vector with the fixed length. Finally, the response result data volume is predicted by a fully connected neural network. To further improve the prediction accuracy, the prediction results of embedded features are combined with the field features and summary features of the query statement to predict the final response data volume by the AutoML model. Findings: Experiments on two public GraphQL API data sets, GitHub and Yelp, show that the accuracy of the proposed approach is 15.85% and 50.31% higher than existing GraphQL response volume prediction approaches based on machine learning techniques, respectively. Originality/value: This paper proposes an approach that combines Code2Vec and AutoML for GraphQL query response data volume prediction with higher accuracy. [ABSTRACT FROM AUTHOR]

We propose the use of a deep learning architecture, called RETINA, to predict multi-alternative, multi-attribute consumer choice from eye movement data. RETINA directly uses the complete time series of raw eye-tracking data from both eyes as input to state-of-the art Transformer and Metric Learning Deep Learning methods. Using the raw data input eliminates the information loss that may result from first calculating fixations, deriving metrics from the fixations data and analysing those metrics, as has been often done in eye movement research, and allows us to apply Deep Learning to eye tracking data sets of the size commonly encountered in academic and applied research. Using a data set with 112 respondents who made choices among four laptops, we show that the proposed architecture outperforms other state-of-the-art machine learning methods (standard BERT, LSTM, AutoML, logistic regression) calibrated on raw data or fixation data. The analysis of partial time and partial data segments reveals the ability of RETINA to predict choice outcomes well before participants reach a decision. Specifically, we find that using a mere 5 s of data, the RETINA architecture achieves a predictive validation accuracy of over 0.7. We provide an assessment of which features of the eye movement data contribute to RETINA's prediction accuracy. We make recommendations on how the proposed deep learning architecture can be used as a basis for future academic research, in particular its application to eye movements collected from front-facing video cameras. [ABSTRACT FROM AUTHOR]

Background: Cancer is positioned as a major disease, particularly for middle-aged people, which remains a global concern that can develop in the form of abnormal growth of body cells at any place in the human body. Cervical cancer, often known as cervix cancer, is cancer present in the female cervix. In the area where the endocervix (upper two-thirds of the cervix) and ectocervix (lower third of the cervix) meet, the majority of cervical cancers begin. Despite an influx of people entering the healthcare industry, the demand for machine learning (ML) specialists has recently outpaced the supply. To close the gap, user-friendly applications, such as H2O, have made significant progress these days. However, traditional ML techniques handle each stage of the process separately; whereas H2O AutoML can automate a major portion of the ML workflow, such as automatic training and tuning of multiple models within a user-defined timeframe. Methods: Thus, novel H2O AutoML with local interpretable model-agnostic explanations (LIME) techniques have been proposed in this research work that enhance the predictability of an ML model in a user-defined timeframe. We herein collected the cervical cancer dataset from the freely available Kaggle repository for our research work. The Stacked Ensembles approach, on the other hand, will automatically train H2O models to create a highly predictive ensemble model that will outperform the AutoML Leaderboard in most instances. The novelty of this research is aimed at training the best model using the AutoML technique that helps in reducing the human effort over traditional ML techniques in less amount of time. Additionally, LIME has been implemented over the H2O AutoML model, to uncover black boxes and to explain every individual prediction in our model. We have evaluated our model performance using the findprediction() function on three different idx values (i.e., 100, 120, and 150) to find the prediction probabilities of two classes for each feature. These experiments have been done in Lenovo core i7 NVidia GeForce 860M GPU laptop in Windows 10 operating system using Python 3.8.3 software on Jupyter 6.4.3 platform. Results: The proposed model resulted in the prediction probabilities depending on the features as 87%, 95%, and 87% for class '0' and 13%, 5%, and 13% for class '1' when idx_value=100, 120, and 150 for the first case; 100% for class '0' and 0% for class '1', when idx_value= 10, 12, and 15 respectively. Additionally, a comparative analysis has been drawn where our proposed model outperforms previous results found in cervical cancer research. [ABSTRACT FROM AUTHOR]

Machine Learning (ML) has evolved beyond being a specialized technique exclusively used by computer scientists. Besides the general ease of use, automated pipelines allow for training sophisticated ML models with minimal knowledge of computer science. In recent years, Automated ML (AutoML) frameworks have become serious competitors for specialized ML models and have even been able to outperform the latter for specific tasks. Moreover, this success is not limited to simple tasks but also complex ones, like tumor segmentation in histopathological tissue, a very time-consuming task requiring years of expertise by medical professionals. Regarding medical image segmentation, the leading AutoML frameworks are nnU-Net and deepflash2. In this work, we begin to compare those two frameworks in the area of histopathological image segmentation. This use case proves especially challenging, as tumor and healthy tissue are often not clearly distinguishable by hard borders but rather through heterogeneous transitions. A dataset of 103 whole-slide images from 56 glioblastoma patients was used for the evaluation. Training and evaluation were run on a notebook with consumer hardware, determining the suitability of the frameworks for their application in clinical scenarios rather than high-performance scenarios in research labs. [ABSTRACT FROM AUTHOR]

Absract: The contemporary field of artificial intelligence has witnessed the emergence of Automated Machine Learning (AutoML) as a noteworthy advancement, promising the delivery of high-performance end-to-end machine learning pipelines with minimal user intervention. While AutoML has exhibited considerable efficacy in various computer vision applications, an unexplored realm remains concerning its application in proximal hyperspectral imaging. The combination of hyperspectral imaging with AutoML for classifying acclimation/stress response levels of horticultural crops is an innovative application nowadays. In this study, PyCaret, an open-source AutoML framework, and PLS1-DA were evaluated for broccoli drought acclimation/stress classification using hyperspectral data. The results revealed that PyCaret and PLS1-DA performed equally well, with PyCaret slightly outperforming PLS1-DA and achieving an accuracy and F1 score of 1.00 both when differentiating between control and drought onset or drought acclimated plants and when differentiating between all three classes. These findings underscore the substantial potential of AutoML and hyperspectral imaging, particularly in tasks related to plants’ water dynamics classification.

Automated machine learning (AutoML) is considered the next advancement in artificial intelligence due to its commitment to delivering high-performance end-to-end machine learning pipelines with minimal user input. Although AutoML has demonstrated significant potential in computer vision tasks, as far as current knowledge extends, there is currently no study that has applied AutoML specifically to hyperspectral imaging for fertilization level classification. To address this information gap, the use of AutoML for classifying fertilization levels using a hyperspectral and CIELAB color space dataset was investigated. A comparative analysis was conducted between the performance of an open-source AutoML framework, PyCaret, and traditional machine learning using the PLS-DA algorithm. PyCaret achieved the highest accuracy (1.00) in classifying different fertilization levels using the hyperspectral dataset, while PLS-DA attained an accuracy of 0.91. However, the CIELAB dataset was not as effective for this classification task, achieving an accuracy of only 0.72. It is worth noting that the hyperspectral dataset outperformed the CIELAB dataset in both AutoML and PLS-DA analyses. Finally, the findings suggest that AutoML holds substantial potential to enhance the use of hyperspectral imaging in agriculture, particularly for fertilization tasks.