Abstract: Abstract Representation learning in neural networks may be implemented with supervised or unsupervised algorithms, distinguished by the availability of instruction. In the sensory cortex, perceptual learning drives neural plasticity 1–13 , but it is not known whether this is due to supervised or unsupervised learning. Here we recorded populations of up to 90,000 neurons simultaneously from the primary visual cortex (V1) and higher visual areas (HVAs) while mice learned multiple tasks, as well as during unrewarded exposure to the same stimuli. Similar to previous studies, we found that neural changes in task mice were correlated with their behavioural learning. However, the neural changes were mostly replicated in mice with unrewarded exposure, suggesting that the changes were in fact due to unsupervised learning. The neural plasticity was highest in the medial HVAs and obeyed visual, rather than spatial, learning rules. In task mice only, we found a ramping reward-prediction signal in anterior HVAs, potentially involved in supervised learning. Our neural results predict that unsupervised learning may accelerate subsequent task learning, a prediction that we validated with behavioural experiments.

Abstract: In credit risk assessment, unsupervised classification techniques can be introduced to reduce human resource expenses and expedite decision-making. Despite the efficacy of unsupervised learning methods in handling unlabeled datasets, their performance remains limited owing to challenges such as imbalanced data, local optima, and parameter adjustment complexities. Thus, this paper introduces a novel hybrid unsupervised classification method, named the two-stage hybrid system with spectral clustering and semi-supervised support vector machine (TSC-SVM), which effectively addresses the unsupervised imbalance problem in credit risk assessment by targeting global optimal solutions. Furthermore, a multi-view combined unsupervised method is designed to thoroughly mine data and enhance the robustness of label predictions. This method mitigates discrepancies in prediction outcomes from three distinct perspectives. The effectiveness, efficiency, and robustness of the proposed TSC-SVM model are demonstrated through various real-world applications. The proposed algorithm is anticipated to expand the customer base for financial institutions while reducing economic losses.

Abstract: Abstract Seismic data recorded at industrial sites contain valuable information on anthropogenic activities. With advances in machine learning and computing power, new opportunities have emerged to explore the seismic wavefield in these complex environments. We applied two unsupervised machine learning algorithms to analyze continuous seismic data collected from an industrial facility in Texas, United States. The Uniform Manifold Approximation and Projection for Dimension Reduction algorithm was used to reduce the dimensionality of the data and generate 2D embeddings. Then, the Hierarchical Density-Based Spatial Clustering of Applications with Noise method was employed to automatically group these embeddings into distinct signal clusters. Our analysis of over 1400 hr (around 59 days) of continuous seismic data revealed five and seven signal clusters at two separate stations. At both stations, we identified clusters associated with background noise and vehicle traffic, with the latter’s temporal patterns aligning closely with the facility’s work schedule. Furthermore, the algorithms detected signal clusters from unknown sources and underline the ability of unsupervised machine learning for uncovering previously unrecognized patterns. Our analysis demonstrates the effectiveness of unsupervised approaches in examining continuous seismic data without requiring prior knowledge or pre-existing labels.

Abstract: Summary The accurate evaluation and prediction of sweet spots are important for realizing scale-effective development of shale oil and gas. Sweet-spot evaluation methods can be categorized into physical model–based methods and data-driven methods. The construction of physical models is usually associated with high labor and material costs, and, coupled with the considerable nonhomogeneity of reservoirs and the involvement of numerous evaluation elements, obtaining physical models with sufficient resolution and accuracy to satisfy actual requirements is challenging. Commonly used labeled data-driven sweet-spot evaluation methods rely heavily on the subjectivity and consistency of user judgment, and most methods require large-scale, high-quality sample training data sets for constrained training, which are rarely applied in geological sweet-spot (GSS)–engineering sweet-spot (ESS) evaluation without labeled samples. In this paper, on the basis of an unsupervised machine learning framework, we developed an integrated GSS-ESS evaluation method for downscaling, clustering, and grading shale horizontal wells via the integration of a priori knowledge and standardized variable-distance evaluation. First, evaluation indices were selected according to the characteristics of the mine data, and the ideal point labels of the evaluation indices were obtained via a priori knowledge and then fused to generate an unsupervised integrated evaluation data set. A standardized variable-distance unsupervised learning clustering model was introduced to calculate the proximity distance of the evaluation indices of GSSs-ESSs to the ideal points of horizontal wells to compress and downscale the evaluation indices into 1D proximity degrees. The evaluation of the proximity distance facilitates the hierarchical assessment of horizontal well sweet spots. A comprehensive evaluation of the sweet spot of fractured horizontal well sections is realized with unsupervised data sets, and the accuracy can reach 86.43% of that obtained by the supervised fuzzy comprehensive evaluation. The results of applying the comprehensive sweet-spot rating to the fracturing segment selection cluster design show that the higher the rating, the better the test production, which indicates the accuracy of this method and provides a new way of thinking for the prediction of sweet spots of unconventional oil and gas resources.

Abstract: Well logging data plays a crucial role in the exploration and extraction of subsurface resources. However, in practical applications, logging data often suffers from missing values or distortions due to geological limitations. To achieve comprehensive subsurface modeling, it is essential to accurately reconstruct this missing data. We assume that logging responses from the same lithology exhibit similar patterns from a petrophysical perspective. Therefore, incorporating lithological information into the logging attribute prediction tasks can enhance the prediction accuracy of the model. We designed a geologically constrained Transformer architecture where the self-attention mechanism enables the model to better understand the relationships between different depth points in the logging data, thereby capturing the complex features of the subsurface structure more accurately. By encoding lithological information as a prior geological constraint and incorporating it along with the logging sequences into the Transformer model, we achieved more accurate predictions for missing logging sequences. To address the challenge of missing lithological data, we introduced the results of Toeplitz inverse covariance-based clustering (TICC) method as a substitute for actual lithological data. The TICC results are used as a geological constraint in the Transformer model to guide the prediction process. Experiments demonstrated that the Transformer combined with TICC technique achieves predictive performance comparable to using actual lithological data, improving the accuracy of logging predictions. This approach provides an effective alternative to practical exploration where real lithological data is not available. Furthermore, we enhanced the predictive capability of the model by designing a regularized loss function that combines the mean-squared error (MSE) with a Gaussian distribution constraint. Application results on field data confirm the reliability and practicality of the geologically constrained Transformer model in accurately predicting acoustic logging.

Abstract: Aiming at the problems of existing time series data clustering methods, such as the lack of similarity metric universality, the influence of dimensional catastrophe, and the limitation of feature expression ability, a time series data clustering method based on unsupervised contrasting learning (UCL-TSC) is proposed. The method first utilizes Residual, TCN, and CNN-TCN to construct multi-view representations of spatial, temporal, and spatial–temporal features of time series data, and adaptively fuses complementary information to enhance feature extraction capabilities. Subsequently, positive and negative sample pairs are constructed based on nearest neighbor and pseudo-clustering label information. Finally, a contrast loss function consisting of feature loss, clustering loss, and a regularization term is designed to facilitate the model in achieving compact intra-cluster and sparse inter-cluster clustering effects in the clustering process. The experimental results on the UCR dataset show that UCL-TSC performs well with respect to several evaluation indexes, such as clustering accuracy, normalized information degree, and purity, and is more effective in learning time series data features and achieving accurate clustering compared to traditional clustering and deep clustering methods.

Abstract: Abstract Well logs, crucial for drilling and post-drilling analysis, provide continuous measurements of subsurface formations as a function of depth. Logging while drilling (LWD) and electrical wireline logs (EWL) are commonly used techniques for well-log acquisition. Both methods are prone to depth measurement errors due to various factors, which need to be aligned to a common depth-scale for subsequent analysis. This study compares two automated machine learning approaches for aligning repeated measurements of the same parameters from LWD and EWL logs of the same well. The first approach is based on supervised learning and the second on unsupervised learning. The supervised approach trains a 1D convolutional neural network (1D CNN) classification model on actual well-log data from the Norwegian North Sea, using LWD-EWL pairs of log slices. A specific depth discrepancy is introduced for each pair, and the logs are divided into various classes based on the depth error between them. The unsupervised method combines autoencoders and K-means clustering to identify potential lithological boundaries in EWL and LWD multiparameter log data. These predicted boundaries are validated by requesting a maximal Pearson correlation. The performance of the classification model is evaluated using metrics such as accuracy, precision, and recall. The optimal number of clusters for K-means clustering is identified using silhouette scores and the elbow method. Depth alignment is verified through visual inspection, correlation analysis, and Euclidean distances between logs. Supervised and unsupervised approaches significantly improve the alignment of various logs, such as bulk density, deep resistivity, sonic compressional, and neutron porosity. Both methods outperform maximization of cross-correlation for specific logs, such as deep resistivity and neutron porosity. These results highlight the potential of machine learning for efficient and accurate depth alignment of well logs, with promising implications for enhancing drilling and post-drilling analysis in the oil and gas industry.

Abstract: Increasing age and age-related health conditions can contribute to a higher risk of crash involvement and more severe physical injuries among older drivers. The current study uses in-vehicle sensing systems to analyze and cluster the overall driving behaviors of older drivers, precisely those aged 65 and older. The primary objective is to develop a framework that can help to observe patterns indicative of distinctive driving styles by clustering and interpreting similar patterns and leveraging self-organizing maps (SOMs) and deep embedded clustering (DEC) methods to reduce the complexity of the sensor data. The research uses two distinct groups of variables, including “speed,” “hard acceleration,” “hard braking” as constant variables, and “rpm,” “throttle positioning,” “fuel level,” “engine level,” and “ambient air temperature” as target variables from the in-vehicle sensor data in understanding driving behaviors and develops models in visualizing and interpreting complex driving patterns by classifying them. The results show that ${5} \times {5}$ grid SOMs have the ability to visualize multiple driving features concurrently, and DEC + K-means and DEC + agglomerative are the well-performed methods for determining the optimal number of clusters to analyze types of driving patterns. The clustering analysis identifies two distinct clusters as the optimal configuration, indicating that the predominant driving behavior among the targeted participants exhibits conservative patterns. The methodologies can be used for other driving features and demographics, making it relevant for broader applications in traffic analysis for understanding and visualizing driving patterns.