Abstract

This paper presents a real-time multi-class classroom behavior detection framework built with YOLOv8 and PyTorch. The system classifies 11+ student behavioral patterns in classroom environments with measurable Precision, Recall, and mAP@50 scores. Grad-CAM explainability layers are integrated to validate model decisions and support responsible AI deployment in educational settings.

Key Contributions

• ›Designed and trained a YOLOv8 multi-class detection model covering 11+ behavioral categories in classroom environments.
• ›Structured and curated a custom annotated dataset with rigorous quality control under CPU-constrained hardware.
• ›Configured reproducible YAML-based training pipelines with systematic hyperparameter optimization.
• ›Achieved mAP@50 improvements through iterative debugging, augmentation strategies, and batch tuning.
• ›Integrated Grad-CAM heatmap visualization for model explainability and responsible AI validation.
• ›Proposed deployment strategies for real-time inference in resource-limited campus settings.

Keywords

Abstract

This paper presents a hybrid CNN-LSTM deep learning architecture for earthquake prediction and the generation of synthetic seismograms. The model leverages convolutional layers for spatial feature extraction from seismic waveform data and LSTM layers for temporal sequence modeling, enabling accurate magnitude prediction and realistic synthetic seismogram synthesis for data augmentation and simulation purposes.

Key Contributions

• ›Designed a hybrid CNN-LSTM architecture that combines convolutional spatial features with LSTM temporal sequence modeling for seismic data analysis.
• ›Developed a pipeline for processing real seismic waveform datasets and training multi-target prediction models for magnitude and location estimation.
• ›Implemented synthetic seismogram generation capabilities to augment training datasets for improved model generalization.
• ›Evaluated model performance against baseline deep learning models on established seismic datasets.
• ›Demonstrated that CNN-LSTM hybrid models outperform standalone CNN or LSTM architectures on temporal seismic prediction tasks.
• ›Published findings in the American Journal of Civil Engineering, October 30, 2025.

Keywords

Abstract

This paper describes the architecture, design decisions, and evaluation of a FastAPI-based resume intelligence system that uses NLP techniques to parse job descriptions and generate tailored career documents. The system achieves 87%+ suitability scoring accuracy and sub-2-second generation latency.

Key Contributions

• ›Described the full architecture of a 6-module FastAPI system for NLP-driven resume generation.
• ›Benchmarked JD skill extraction accuracy against manually labeled datasets (87%+ suitability match).
• ›Evaluated PDF/DOCX generation latency and formatting consistency across diverse input lengths.
• ›Compared regex-based NLP heuristics against spaCy NER models for skill extraction precision.
• ›Proposed future directions including LLM-powered summary rewriting and ATS score prediction.

Keywords