Automatic Data Processing (United States)

The paper proposes concept of managing territorial-production complexes (TPC) based on Continuous Intelligence (CI) principles. The developed management system fundamentally differs from traditional cybernetic models by shifting from discrete analysis to continuous monitoring and dynamic adaptation of control inputs in real-time. As a technological foundation, the authors use the DATA-CI platform implementing the method of object relations for integrating existing enterprise IT systems (Process Automation Systems, Manufacturing Execution Systems, Enterprise Resource Planning, Business Intelligence) without replacing them. The authors have tested the platform on a TPC economy model encompassing twelve production stages, demonstrating capabilities for real-time visualization of system states through mimetic diagrams, effective collection, and storage of data on all production processes. The obtained results confirm the advantages of the proposed approach, including reduced decision-making lag, improved forecasting accuracy, and comprehensive management of the TPC as a unified system. Further development prospects involve expanding the functionality of the platform and adapting the proposed concept to various types of territorial-production complexes.

The growing volume, speed, and diversity of contemporary data streams require adaptive feature engineering pipelines that are able to learn adaptively on dynamic settings. To minimize the time spent in feature extraction, the proposed paper offers a Self-Learning AI Pipeline of Dynamic Big Data Feature Engineering, where all three-reinforcement learning, meta-learning, and automated data preprocessing mechanisms are integrated to optimize feature extraction in real time. This framework keeps track of the incoming data distributions, identifies concept drifts, and adjusts transformation strategies with feedback-based reinforcement agent. By using a hybrid design of a stream-based data ingestion, autoencoder-driven feature selection, hyperparameter tuning by Bayesian optimization, the adaptability is ensured and manual intervention is kept to a minimum. Large-scale benchmark datasets show that experiments with these models achieve high model accuracy, feature relevance and lower computational costs than traditional static pipelines. The proposed system accelerates predictive tasks of downstream the adaptation to a data drift by up to 30 percent, and the performance in the system increases by 1520 percent. This study forms a basis to autonomous and self-optimizing ecosystems of feature engineering in big data analytics, enabling scalable and intelligent data pipelines to support next-generation AI applications.

Since Nigeria implemented its new cashless policy and introduced new naira notes, there has been a dramatic rise in ATM card fraud. The widespread use of credit and debit cards makes them easy targets for fraudsters, which worries both financial institutions and their customers. This, in turn, slows down economic growth and breeds mistrust among consumers. There have been multiple efforts to study the problems and proposed remedies using various methods, including traditional methodologies, machine learning, enhanced firewall security, probability, etc. However, data collecting and fraud investigation have both shown that it is difficult to detect and prevent ATM card fraud for the following reasons: most conventional investigators lack the expertise to detect it due to its infrequency and parallel nature, and there is a lack of knowledge about the access point where ATM fraud is committed. In the study, to suggested two methods for detecting ATM fraud: one for static contexts and another for streaming scenarios. The primary objective of this study is to create an intelligent agent-based model that can detect ATM card frauds with high performance. Basically, the model involves 3-factor authentication process to determine if a user should be granted access or otherwise, deny access to a bank ATM, these includes; Password module, AIAgent spending patter detection and lastly Biometric modules authentication. Optimized Bi-Directional Long Short-Term Memory model were applied to historical log files of the ATM cards usage patterns to devise an intelligent data mining classification model. The weight of the proposed model is optimally selected by Mapping based Cuttlefish optimization that improves the classification accuracy. This model was evaluated using confusion matrix with the dataset from the bank. This will help immensely in providing a way forward in curbing the high rate of ATM card fraud in banking sector.

     This paper introduces a reactive navigation strategy for wheeled mobile robots, utilizing type-2 fuzzy logic to manage uncertainty in dynamic environments. The approach incorporates two distinct type-2 fuzzy logic controllers, each tailored to address specific challenges in navigation. The first controller focuses on steering the robot toward its target by continuously adjusting its path in response to changing conditions. The second controller specializes in obstacle avoidance, enabling the robot to detect and maneuver around obstacles it encounters during its journey.      To evaluate the performance of the system, numerical simulations are carried out across various scenarios, including dynamic and cluttered environments, to demonstrate its robustness. Additionally, the results of the type-2 fuzzy logic approach are compared with conventional navigation techniques, such as rule-based or model-based methods. The comparison underscores the system’s greater adaptability and resilience. The study concludes that type-2 fuzzy logic provides an effective and flexible solution, significantly improving both path planning and real-time decision-making in unpredictable and complex environments.

The field of autonomous vehicles (AVs) predominantly leverages multi-modal integration of LiDAR and camera data to achieve better performance compared to using a single modality. However, the fusion process encounters challenges in detecting distant objects due to the disparity between the high resolution of cameras and the sparse data from LiDAR. Insufficient integration of global perspectives with local-level details results in sub-optimal fusion performance.To address this issue, we have developed an innovative two-stage fusion process called Quantum Inverse Contextual Vision Transformers (Q-ICVT). This approach leverages adiabatic computing in quantum concepts to create a novel reversible vision transformer known as the Global Adiabatic Transformer (GAT). GAT aggregates sparse LiDAR features with semantic features in dense images for cross-modal integration in a global form. Additionally, the Sparse Expert of Local Fusion (SELF) module maps the sparse LiDAR 3D proposals and encodes position information of the raw point cloud onto the dense camera feature space using a gating point fusion approach. Our experiments show that Q-ICVT achieves an mAPH of 82.54 for L2 difficulties on the Waymo dataset, improving by 1.88% over current state-of-the-art fusion methods. We also analyze GAT and SELF in ablation studies to highlight the impact of Q-ICVT. Our code is available at https://github.com/sanjay-810/Qicvt