Raw Input Research Articles

End-to-End Autonomous Driving: Challenges and Frontiers.

The autonomous driving community has witnessed a rapid growth in approaches that embrace an end-to-end algorithm framework, utilizing raw sensor input to generate vehicle motion plans, instead of concentrating on individual tasks such as detection and motion prediction. End-to-end systems, in comparison to modular pipelines, benefit from joint feature optimization for perception and planning. This field has flourished due to the availability of large-scale datasets, closed-loop evaluation, and the increasing need for autonomous driving algorithms to perform effectively in challenging scenarios. In this survey, we provide a comprehensive analysis of more than 270 papers, covering the motivation, roadmap, methodology, challenges, and future trends in end-to-end autonomous driving. We delve into several critical challenges, including multi-modality, interpretability, causal confusion, robustness, and world models, amongst others. Additionally, we discuss current advancements in foundation models and visual pre-training, as well as how to incorporate these techniques within the end-to-end driving framework.

Finger Vein Detection using Deep Learning

The deep learning is a part of machine learning that uses neural networks with many layers (deep architecture) to fit the intricate patterns. Deep learning has shown remarkable performance in image classification mainly due to its ability in learning and extracting features from the raw inputs. In this research study, some techniques of image preprocessing were described and utilized to machine learning algorithm for image classification. The purpose was to promote the specific preprocessing techniques on the impact, which are commonly used in classification model of images, relating to time and precision improvement. The techniques that are applied in this documentation include; image enhancement, binarization, transformation, edge detection and scaling. The present research study therefore comes handy in trying to establish the significance of enhancing statistic and the impact on the performance of the model. The following preprocessing methods have been used because the image datasets need preprocessing to be input for the machine learning models for better image classification

Text-independent speaker identification using modified SincNet with robust features from suitable acoustic region and appropriate optimizer for raw audio analysis

Speaker identification is a method of identifying an individual from a set of speakers, and text-independent speaker identification systems allow speakers to utter any phrase without any constraints. This study is focused on raw audio analysis as phase, fine-grained frequency patterns, timing cues, and other minute characteristics are preserved when raw waveforms are processed as compared to handcrafted features like Mel-Frequency Cepstral Coefficients (MFCC) and visual representation of audio-like spectrogram. Due to the depth of information, which includes variations in speech rhythm, pitch, and vocal tract shape, it is beneficial for identifying speakers. The deep learning architecture known as SincNet has gained popularity in speaker identification because of its parametric Sinc functions that allow it to operate directly on the raw audio input. In this paper, we have considered SincNet as the baseline model for speaker identification. The effect of proper speech boundary detection, including high-level features and effective optimizer selection are analysed. The precise identification of the signal start and terminus point is important for eliminating the redundant non-speech regions. We have included endpoint detection module as a pre-processing step in the system. Proper feature extraction and selection are crucial to the model's success. To extract more abstract features from the data, we have added more convolution layers to the original SincNet model. Further, we investigated the hyperparameter tuning protocol's sensitivity to the optimizer and selected the suitable optimizer for raw audio analysis. With all the modifications in the system architecture, we are able to archive improvements of 12.76 %, 13.33 %, and 13.39 % respectively for training, validation, and testing over the original SincNet model. In terms of validation loss, our proposed approach attains 0.35 in comparison to the original SincNet loss of 1.02. With this significant improvement, the total training time is marginally increased by 20 minutes for our proposed model. We have performed our investigation on the LibriSpeech dataset to check the effectiveness of our proposed system in comparison to the other model..

GaussianHeads: End-to-End Learning of Drivable Gaussian Head Avatars from Coarse-to-fine Representations

Real-time rendering of human head avatars is a cornerstone of many computer graphics applications, such as augmented reality, video games, and films, to name a few. Recent approaches address this challenge with computationally efficient geometry primitives in a carefully calibrated multi-view setup. Albeit producing photorealistic head renderings, they often fail to represent complex motion changes, such as the mouth interior and strongly varying head poses. We propose a new method to generate highly dynamic and deformable human head avatars from multi-view imagery in real time. At the core of our method is a hierarchical representation of head models that can capture the complex dynamics of facial expressions and head movements. First, with rich facial features extracted from raw input frames, we learn to deform the coarse facial geometry of the template mesh. We then initialize 3D Gaussians on the deformed surface and refine their positions in a fine step. We train this coarse-to-fine facial avatar model along with the head pose as learnable parameters in an end-to-end framework. This enables not only controllable facial animation via video inputs but also high-fidelity novel view synthesis of challenging facial expressions, such as tongue deformations and fine-grained teeth structure under large motion changes. Moreover, it encourages the learned head avatar to generalize towards new facial expressions and head poses at inference time. We demonstrate the performance of our method with comparisons against the related methods on different datasets, spanning challenging facial expression sequences across multiple identities. We also show the potential application of our approach by demonstrating a cross-identity facial performance transfer application. We make the code available on our project page.

The Combination of Topological Data Analysis and Mathematical Modeling Improves Sleep Stage Prediction From Consumer-Grade Wearables.

Wearable devices have become commonplace tools for tracking behavioral and physiological parameters in real-world settings. Nonetheless, the practical utility of these data for clinical and research applications, such as sleep analysis, is hindered by their noisy, large-scale, and multidimensional characteristics. Here, we develop a neural network algorithm that predicts sleep stages by tracking topological features (TFs) of wearable data and model-driven clock proxies (CPs) reflecting the circadian propensity for sleep. To evaluate its accuracy, we apply it to motion and heart rate data from the Apple Watch worn by young subjects undergoing polysomnography (PSG) and compare the predicted sleep stages with the corresponding ground truth PSG records. The neural network that includes TFs and CPs along with raw wearable data as inputs shows improved performance in classifying Wake/REM/NREM sleep. For example, it shows significant improvements in identifying REM and NREM sleep (AUROC/AUPRC improvements >13% and REM/NREM accuracy improvement of 12%) compared with the neural network using only raw data inputs. We find that this improvement is mainly attributed to the heart rate TFs. To further validate our algorithm on a different population, we test it on elderly subjects from the Multi-ethnic Study of Atherosclerosis cohort. This confirms that TFs and CPs contribute to the improvements in Wake/REM/NREM classification. We next compare the performance of our algorithm with previous state-of-the-art wearable-based sleep scoring algorithms and find that our algorithm outperforms them within and across different populations. This study demonstrates the benefits of combining topological data analysis and mathematical modeling to extract hidden inputs of neural networks from puzzling wearable data.

The Computational Effect And Hyperparameters Tuning Of Deep Convolutional Layer Depth Of High- Ranking Tuberculosis Detection Models

In the past decade, artificial intelligence (AI) solutions have seen widespread application in medical fields, which include automated detection of breast cancer, brain tumors, physiological monitoring and detection of lung diseases such as pneumonia from chest X-Rays (CXRs). Machine learning, a subset of AI, empowers computers to learn autonomously, without direct human programming, by extracting patterns (feature extraction) from data (images). Deep learning, a specialized branch of machine learning, employs multiple convolutional layers to extract complex features from raw input data. This article examines the impact of varying the convolutional layers in deep learning models on their efficiency, focusing on algorithm complexity and parameter counts. We discuss theoretical foundations, and relevant factors affecting efficiency, and analyze algorithm complexity of high-ranking models developed for detecting tuberculosis from chest x-ray (CXR) using convolutional neural networks. The number of convolutional layers significantly influences model efficiency, affecting both performance and computational complexity. We could conclude practically that optimal layer depth balances model efficiency, accuracy and resource utilization. This assertion was reached by developing a model with fewer convolutional layer depth using the ResNet18 architecture. The parameters count of the ResNet18 model developed was compared with other model developed to detect tuberculosis from chest X-Ray images. The result of this comparison proved that fewer layer depth with the right hyperparameter tuning can produce a better and more efficient deep learning solutions to societal problems. This article also gives insights for future research and practical applications which includes the exploration of adaptive architectures that can dynamically adjust their depth based on the complexity of the task and available resources. Further research can also scientifically probe other methods of reducing the computational overhead of deep neural networks while maintaining priority for high computational performance, model scalability and efficiency.

T-KAER: Transparency-enhanced Knowledge-Augmented Entity Resolution Framework

Entity resolution (ER) is the process of determining whether two representations refer to the same real-world entity and plays a crucial role in data curation and data cleaning. Recent studies have introduced the KAER framework, aiming to improve pre-trained language models by augmenting external knowledge. However, identifying and documenting the external knowledge that is being augmented and understanding its contribution to the model's predictions have received little to no attention in the research community. This paper addresses this gap by introducing T-KAER, the Transparency-enhanced Knowledge-Augmented Entity Resolution framework. To enhance transparency, three Transparency-related Questions (T-Qs) have been proposed: T-Q(1): What is the experimental process for matching results based on data inputs? T-Q(2): Which semantic information does KAER augment in the raw data inputs? T-Q(3): Which semantic information of the augmented data inputs influences the predictions? To address the T-Qs, T-KAER is designed to improve transparency by documenting the entity resolution processes in log files. In experiments, a citation dataset is used to demonstrate the transparency components of T-KAER. This demonstration showcases how T-KAER facilitates error analysis from both quantitative and qualitative perspectives, providing evidence on `"what" semantic information is augmented and ``"why" the augmented knowledge influences predictions differently.

Firm size and enterprise productivity in Nigeria: Evidence from firm-level data

Anchored on the economies of scale of production theory, this study utilized Nigeria’s firm-level enterprise survey data of the World Bank collected through stratified random sampling of 2 676 firms and face-to-face interviews with the application of the multinomial logit model to examine how enterprise productivity influences the size of firms in Nigeria. It was found that raising enterprise productivity relates to about 0.0009261 insignificant fall in the relative log odds of running micro-sized firms, about 0.010299 significant rise in relative log odds of having medium-sized firms, and about 0.0201428 significant encouragement in relative log odds of running large-sized enterprises/firms when related with small-sized enterprises/firms. It is recommended that governments at all levels (state, federal, and local), should encourage micro-sized firms in a bid to make them increase their productivity level. This encouragement can come in the form of providing increased access to credit, the provision of raw material inputs, and constant electricity supplies. The original contribution of this research work is hinged on its empirical contribution in the study area since there is dearth of literature in the study area as no study has looked at firm size and enterprise productivity in Nigeria using evidence from firm-level data.

Sustainable Production of Biomass-Derived Graphite and Graphene Conductive Inks from Biochar.

Graphite is a commonly used raw material across many industries and the demand for high-quality graphite has been increasing in recent years, especially as a primary component for lithium-ion batteries. However, graphite production is currently limited by production shortages, uneven geographical distribution, and significant environmental impacts incurred from conventional processing. Here, an efficient method of synthesizing biomass-derived graphite from biochar is presented as a sustainable alternative to natural and synthetic graphite. The resulting bio-graphite equals or exceeds quantitative quality metrics of spheroidized natural graphite, achieving a Raman ID/IG ratio of 0.051 and crystallite size parallel to the graphene layers (La) of 2.08 µm. This bio-graphite is directly applied as a raw input to liquid-phase exfoliation of graphene for the scalable production of conductive inks. The spin-coated films from the bio-graphene ink exhibit the highest conductivity among all biomass-derived graphene or carbon materials, reaching 3.58 ± 0.16 × 104 Sm-1. Life cycle assessment demonstrates that this bio-graphite requires less fossil fuel and produces reduced greenhouse gas emissions compared to incumbent methods for natural, synthesized, and other bio-derived graphitic materials. This work thus offers a sustainable, locally adaptable solution for producing state-of-the-art graphite that is suitable for bio-graphene and other high-value products.

Knowledge abstraction and filtering based federated learning over heterogeneous data views in healthcare

Robust data privacy regulations hinder the exchange of healthcare data among institutions, crucial for global insights and developing generalised clinical models. Federated learning (FL) is ideal for training global models using datasets from different institutions without compromising privacy. However, disparities in electronic healthcare records (EHRs) lead to inconsistencies in ML-ready data views, making FL challenging without extensive preprocessing and information loss. These differences arise from variations in services, care standards, and record-keeping practices. This paper addresses data view heterogeneity by introducing a knowledge abstraction and filtering-based FL framework that allows FL over heterogeneous data views without manual alignment or information loss. The knowledge abstraction and filtering mechanism maps raw input representations to a unified, semantically rich shared space for effective global model training. Experiments on three healthcare datasets demonstrate the framework’s effectiveness in overcoming data view heterogeneity and facilitating information sharing in a federated setup.

Highfield imaging of the subgenual anterior cingulate cortex in uni- and bipolar depression.

The subgenual Anterior Cingulate Cortex (sgACC), as a part of the Anterior Cingulate Cortex and the limbic system plays a crucial role in mood regulation. Previous structural and functional brain imaging studies of the sgACC have revealed alterations of Gray Matter (GM) volumes and Blood Oxygenation Level Dependent signals (BOLD) in patients with Major Depressive Disorder (MDD) and Bipolar Disorder (BD), suggesting potential biomarker traits for affective disorders. In this study we investigated the gray matter volume of the sgACC in 3 different patient groups: 40 MDD patients, of which 20 were medicated (MDDm) and 20 were unmedicated (MDDu), and 21 medicated BD patients, and compared them with 23 healthy volunteers. We examined GM volume alteration using high-resolution 7T Magnetic Resonance Imaging (MRI) which produced quantitative maps of the spin-lattice relaxation time (T1). T1 maps provide high contrast between gray and white matter, and at 7 Tesla voxels with submillimeter resolution can be acquired in a reasonable scan time. We developed a semi-automatic segmentation protocol based on refined landmarks derived from previous volumetric studies using quantitative T1 maps as raw input data for automatic tissue segmentation of GM, WM and CSF (cerebrospinal fluid) tissue. The sgACC ROI was then superimposed on these tissue probability maps and traced manually by two independent raters (F.B., M.M.) following our semi-automatic segmentation protocol. Interrater reliability was calculated on a subset of 10 brain scans for each hemisphere, showing an Intra-Class Correlation coefficient (ICC) r = 0.96 for left sgACC and r = 0.84 for right sgACC respectively. In summary, we have developed a reproducible and reliable semi-automatic segmentation protocol to measure gray matter volume in the sgACC. Based on previous findings from meta-analyses on morphometric studies of the sgACC, we hypothesized that patients with MDD would have lower gray matter sgACC volumes compared to healthy subjects. Post-hoc analysis revealed smaller subgenual volumes for the left hemisphere in both the medicated (MDDm) and non-medicated (MDDu) group versus healthy controls (p = .001, p = .008) respectively. For the right hemisphere, the (MDDu) and BD group exhibited significantly lower subgenual volumes than healthy controls (p < .001, p = .004) respectively. To our knowledge, this is the first morphometric MRI study using T1 maps obtained in high-resolution 7 Tesla MRI to compare MDD and BD patients with healthy controls.

A Deep Learning-Driven Sampling Technique to Explore the Phase Space of an RNA Stem-Loop.

The folding and unfolding of RNA stem-loops are critical biological processes; however, their computational studies are often hampered by the ruggedness of their folding landscape, necessitating long simulation times at the atomistic scale. Here, we adapted DeepDriveMD (DDMD), an advanced deep learning-driven sampling technique originally developed for protein folding, to address the challenges of RNA stem-loop folding. Although tempering- and order parameter-based techniques are commonly used for similar rare-event problems, the computational costs or the need for a priori knowledge about the system often present a challenge in their effective use. DDMD overcomes these challenges by adaptively learning from an ensemble of running MD simulations using generic contact maps as the raw input. DeepDriveMD enables on-the-fly learning of a low-dimensional latent representation and guides the simulation toward the undersampled regions while optimizing the resources to explore the relevant parts of the phase space. We showed that DDMD estimates the free energy landscape of the RNA stem-loop reasonably well at room temperature. Our simulation framework runs at a constant temperature without external biasing potential, hence preserving the information on transition rates, with a computational cost much lower than that of the simulations performed with external biasing potentials. We also introduced a reweighting strategy for obtaining unbiased free energy surfaces and presented a qualitative analysis of the latent space. This analysis showed that the latent space captures the relevant slow degrees of freedom for the RNA folding problem of interest. Finally, throughout the manuscript, we outlined how different parameters are selected and optimized to adapt DDMD for this system. We believe this compendium of decision-making processes will help new users adapt this technique for the rare-event sampling problems of their interest.

Classification of the L-, H-mode, and plasma-free state: Convolutional neural networks and variational autoencoders on the edge reflectometer for KSTAR.

Classifying and monitoring the L-, H-mode, and plasma-free state are essential for the stable operational control of tokamaks. Edge reflectometry measures plasma density profiles, but the large volume of data and complexity in reconstruction pose significant challenges. There is a need for efficient methods to analyze complex reflectometer data in real-time, which can be addressed using advanced computational techniques. Here, we show that machine learning (ML) techniques can classify discharge states using raw signal data from an edge reflectometer installed on the Korea Superconducting Tokamak Advanced Research. The deep convolutional neural network models achieved classification accuracy of up to 99% when using 2D spectrogram inputs, demonstrating a significant improvement over 1D raw signal inputs. Additionally, the variational autoencoder model effectively clustered the discharge states in the latent space without any label information, further validating the model's capability to classify discharge states. These results suggest that the ML model can effectively handle the complexity of reflectometer data and accurately classify plasma discharge states. This approach not only facilitates real-time diagnosis but also reduces the need for manual data processing.

Automated Provenance-Based Screening of ML Data Preparation Pipelines

SummarySoftware systems that learn from data via machine learning (ML) are being deployed in increasing numbers in real world application scenarios. These ML applications contain complex data preparation pipelines, which take several raw inputs, integrate, filter and encode them to produce the input data for model training. This is in stark contrast to academic studies and benchmarks, which typically work with static, already prepared datasets. It is a difficult and tedious task to ensure at development time that the data preparation pipelines for such ML applications adhere to sound experimentation practices and compliance requirements. Identifying potential correctness issues currently requires a high degree of discipline, knowledge, and time from data scientists, and they often only implement one-off solutions, based on specialised frameworks that are incompatible with the rest of the data science ecosystem.We discuss how to model data preparation pipelines as dataflow computations from relational inputs to matrix outputs, and propose techniques that use record-level provenance to automatically screen these pipelines for many common correctness issues (e.g., data leakage between train and test data). We design a prototypical system to screen such data preparation pipelines and furthermore enable the automatic computation of important metadata such as group fairness metrics. We discuss how to extract the semantics and the data provenance of common artifacts in supervised learning tasks and evaluate our system on several example pipelines with real-world data.

Automatically learning linguistic structures for entity relation extraction

A sentence is composed of linguistically linked units, such as words or phrases. The dependencies between them compose the linguistic structures of a sentence, which indicates the meanings of linguistic units and encodes the syntactic or semantic relationships between them. Therefore, it is important to learn the linguistic structures of a sentence for entity relation extraction or other natural language processing (NLP) tasks. In related works, manual rules or dependency trees are usually adopted to capture the linguistic structures. These methods heavily depend on prior knowledge or external toolkits. In this paper, we introduce a Supervised Graph Autoencoder Network (SGAN) model to automatically learn the linguistic structures of a sentence. Unlike traditional graph neural networks that use a fixed adjacency matrix initialized with prior knowledge, the SGAN model contains a learnable adjacency matrix that is dynamically tuned by a task-relevant learning objective. It can automatically learn linguistic structures from raw input sentences. After being evaluated on seven public datasets, the SGAN achieves state-of-the-art (SOTA) performance, outperforming all compared models. The results show that automatically learned linguistic structures have better performance than manually designed linguistic patterns. It exhibits great potential for supporting entity relation extraction and other NLP tasks.

Optimization and characterization of boric acid powder mixed electrical discharge machining of Ti-6Al-4V ELI

The powder mixed electrical discharge machining (PMEDM) process is used for enhancing machining efficiency and improving the surface characteristics of the workpiece compared to normal EDM. The present study evaluates the effect of boric acid powder mixed in deionized water in PMEDM of Ti-6Al-4V ELI (extra low interstitial). The PMEDM process is assessed by evaluating material removal rate (MRR) and surface roughness (SR) at varying powder concentration (PC), current and pulse duration ( Ton). Boric acid PMEDM exhibits advantageous for both MRR and SR, where MRR increases and SR decreases with an increase in PC. To simultaneously achieve higher MRR and lower SR, the grey-fuzzy logic method is employed. Initially, the raw input data is normalized into grey numbers and a single index is obtained through grey relational analysis. Subsequently, triangular membership functions are assigned to the inputs and output. Fuzzy logic rules are then formulated, and grey fuzzy relational grade (GFRG) is calculated for each condition. The optimum condition (GFRG = 0.787) is attained at 15 g/l PC, 9 A current and 106 µs Ton, resulting in a 63% improvement compared to the initial condition. At the optimum condition, MRR increased by 145% and SR decreased by 29% compared to the initial condition. The effect of boric acid in PMEDM can be visualized through the formation of smooth surfaces, exhibiting reduced SR, smaller and shallower craters, and fewer debris and globules. Additionally, the boric acid PMEDM experiments resulted in the formation of a hard TiB phase on Ti-6Al-4V, thus improving its wear resistance.

Machine cross-domain remaining useful life prediction via contrastive adversarial variational recurrent method

The performance degradation process of the machine is non-stationary which varies with the operating environment and is reflected as a temporal shift phenomenon. Models or methods that assume the test and training sets have the same distribution will no longer be suitable for solving problems where domain shifts exist. This brings new challenges for accurately evaluating the machine’s remaining useful life (RUL). This paper studies a contrastive adversarial variational recurrent method for machine RUL prediction under different service conditions. In this new approach, the variational recurrent networks are developed to extract distribution features and latent spaces of raw data, meanwhile, the adversarial strategies are applied to enable the learning of domain-invariant features to make the model achieve cross-domain task processing. To supervise the learning process of the model to learn more mutual information between the extracted features and the raw input data, a contrastive loss is also introduced in the proposed method. Sufficient experiments were conducted to verify the feasibility and superiority of the suggested approach, including 12 sets of cross-scenario tests on the turbofan engine dataset C-MAPSS from NASA. Experimental findings confirm that the proposed method performs satisfactorily and competitively even compared to current state-of-the-art works.

Timely ICU Outcome Prediction Utilizing Stochastic Signal Analysis and Machine Learning Techniques with Readily Available Vital Sign Data.

The ICU is a specialized hospital department that offers critical care to patients at high risk. The massive burden of ICU-requiring care requires accurate and timely ICU outcome predictions for alleviating the economic and healthcare burdens imposed by critical care needs. Existing research faces challenges such as feature extraction difficulties, low accuracy, and resource-intensive features. Some studies have explored deep learning models that utilize raw clinical inputs. However, these models are considered non-interpretable black boxes, which prevents their wide application. The objective of the study is to develop a new method using stochastic signal analysis and machine learning techniques to effectively extract features with strong predictive power from ICU patients' real-time time series of vital signs for accurate and timely ICU outcome prediction. The results show the proposed method extracted meaningful features and outperforms baseline methods, including APACHE IV (AUC = 0.750), deep learning-based models (AUC = 0.732, 0.712, 0.698, 0.722), and statistical feature classification methods (AUC = 0.765) by a large margin (AUC = 0.869). The proposed method has clinical, management, and administrative implications since it enables healthcare professionals to identify deviations from prognostications timely and accurately and, therefore, to conduct proper interventions.

Drawback And Its Contributions To Exports: Analysis Of The Special Regime In Brazilian Companies In The Piracicaba/Sp Region

Background: The drawback is a special customs regime designed to allow companies to exempt or defer taxes and fees on imported raw materials, inputs, packaging materials, and intermediate products that are used in the production of goods intended for export. This mechanism helps lower the cost of finished products, thereby enhancing their competitiveness in the global market. As an export incentive, the drawback plays a significant role in alleviating the tax burden on manufacturers. This study sought to explore the concept of drawback, its different modalities, and the key benefits it offers. Materials and Methods: The research methodology included a literature review and a case study involving two companies located in the Piracicaba/SP region that utilize this customs regime. Data was gathered through a questionnaire with open-ended questions directed at the Foreign Trade managers of these companies. Results: The findings reveal that the companies studied leverage the drawback regime to significantly reduce the prices of their export products, taking advantage of the tax benefits to lower costs and enhance product competitiveness Conclusion: Both theory and practice show that this special customs regime serves as a stimulus to imports and exports, thus enabling the growth of companies in the international market. It can be said that it is an important mechanism for promoting the internationalization process of companies.

Learning the feature distribution similarities for online time series anomaly detection

Identifying anomalies in multi-dimensional sequential data is crucial for ensuring optimal performance across various domains and in large-scale systems. Traditional contrastive methods utilize feature similarity between different features extracted from multidimensional raw inputs as an indicator of anomaly severity. However, the complex objective functions and meticulously designed modules of these methods often lead to efficiency issues and a lack of interpretability. Our study introduces a structural framework called SimDetector, which is a Local–Global Multi-Scale Similarity Contrast network. Specifically, the restructured and enhanced GRU module extracts more generalized local features, including long-term cyclical trends. The multi-scale sparse attention module efficiently extracts multi-scale global features with pattern information. Additionally, we modified the KL divergence to suit the characteristics of time series anomaly detection, proposing a symmetric absolute KL divergence that focuses more on overall distribution differences. The proposed method achieves results that surpass or approach the State-of-the-Art (SOTA) on multiple real-world datasets and synthetic datasets, while also significantly reducing Multiply-Accumulate Operations (MACs) and memory usage.