A Survey on Retrieval-Augmented and Transformer-Based Approaches for Legal AI: Grounding, Reasoning, and Multilingual Outcome Prediction
Legal artificial intelligence (AI) systems increasingly draw on retrieval-augmented generation (RAG), transformerbased embeddings, multi-stage structured reasoning, and ensemble learning to support tasks such as legal question answering, legal judgment prediction, and multilingual outcome prediction. Hallucination, lim...