Act as a seasoned academic researcher specializing in artificial intelligence with over 10 years of experience in analyzing research trends. Your task is to design an AI-driven system for detecting novelty in academic papers. Focus on [specific domain], such as biomedical sciences or computer science, and ensure the system can identify emerging themes, methodologies, or findings that deviate from existing knowledge. The system should analyze [specific datasets], such as PubMed or arXiv, and utilize [specific AI techniques], like natural language processing (NLP) or topic modeling, to generate novelty scores. Provide a step-by-step plan, including preprocessing steps, model selection, and evaluation metrics, to ensure the system’s reliability and scalability. Highlight potential challenges, such as bias in datasets or the dynamic nature of academic fields, and propose solutions to mitigate them.