Articles

In the present research work, an attempt is made to enhance the California Bearing Ratio (CBR) values of subgrade soils by using industrial waste materials such as fly ash (FA) and rice husk ash (RHA) with Artificial Neural Network (ANN)–based prediction to strengthen the reliability of pavement design decisions. Soil samples collected from varied locations were stabilized with 5%, 10%, and 15% FA and RHA, and evaluated through compaction and California Bearing Ratio (CBR) tests in both soaked and unsoaked conditions. To complement laboratory analysis, a classification-based ANN model was developed using the MATLAB Neural Network Toolbox, incorporating stabilizer type, dosage, MDD, OMC, and CBR as input parameters. The ANN predicted CBR category, cost reduction, and pavement thickness, while prediction accuracy was assessed using confusion matrices and associated performance metrics. Results showed significant improvement in CBR values with increasing stabilizer content, particularly with FA. The ANN model achieved high prediction accuracy, validating the consistency of experimental outcomes. The study demonstrates that combining experimental and ANN-based approaches provides a robust decision-support framework for soil stabilization and flexible pavement design.

Permanent Magnet Synchronous Motors (PMSMs) are widely used in high-performance electric drives due to their high efficiency, torque density, and fast dynamic response. This paper presents a comparative analysis of two advanced torque and current control strategies—Field Oriented Control (FOC) and Direct Predictive Torque Control (DPTC)—applied to PMSM drives. The study evaluates their relative performance across key dimensions including transient response, steady-state characteristics, torque and current ripple, switching frequency behavior, computational complexity, and robustness to parameter variations. Drawing from literature spanning 2010–2025 and supported by simulation-based observations, results show that while FOC provides smooth steady-state performance and ease of implementation, predictive torque control methods exhibit superior transient performance and greater flexibility under dynamic operating conditions. The paper concludes with practical guidelines for selecting control strategies based on application-specific constraints such as switching losses, real-time computational limits, and desired dynamic behavior.

This paper proposes a grammatical error correction (GEC) framework that combines effectively the strengths of reinforcement learning with those of transform-based language models. The uniqueness of this research effort specifically resides in using a policy-level Q-learning mechanism to adaptively re-rank potential corrections from a generative model, going beyond traditional approaches that rely solely on T5 for error corrections in text, for example, or even rely on a mere re-rank with a model like GPT2. This proposed scheme uses a fine-tuned version of T5 for error corrections as a generative model that can provide a list of potential corrections, and a separate GPT2 model that will provide an implicit reward as a judge of grammatical fluency. Subsequently, an agent algorithm will help develop an optimal policy that fills up a Q-table that associates error states with superior actions for error corrections. To serve as a preliminary indication of feasibility, this research effort proposes a qualitative analysis with a targeted data set due to inherent limitations in scope.