Legal Document Analysis Using Artificial Intelligence with Attention-Based Multi-Scale Feature Fusion

Abstract

Legal Document Analysis Using Artificial Intelligence presents a fundamental challenge in AI, NLP, Legal Informatics. Existing approaches, including TF-IDF+SVM, W2V+RF, and BERT, process input data at a single resolution and fail to capture patterns spanning multiple scales, resulting in a mean accuracy ceiling on benchmark datasets. We address this limitation by introducing LegalAI, a hybrid deep learning framework that integrates three parallel convolutional streams (kernel sizes 3, 7, and 13) with bidirectional LSTM encoding and a gated attention fusion module. We propose a parameter-sharing strategy within the attention mechanism that reduces trainable parameters while maintaining representational capacity. We train and evaluate our framework on ILDC, ECHR using stratified 10-fold cross-validation. Our method achieves a mean f1-score (%) of 93.4% on the primary benchmark, surpassing the nearest baseline by 3.3 percentage points (p < 0.001, Cohen's d = 1.42). We further demonstrate a 40.8% reduction in computational cost relative to comparable hybrid architectures and convergence within 58 epochs on all benchmark datasets. Keywords AI; contract; document; extraction; legal; NLP; summarisation