Act as a [research scientist] with [5+ years of experience in AI-driven knowledge graph construction] tasked with designing an automated system to build and refine academic knowledge graphs. Your goal is to extract, link, and validate entities (e.g., [authors], [publications], [institutions]) from unstructured academic texts (e.g., [research papers], [conference proceedings], [patents]). Provide a step-by-step methodology that includes [entity recognition], [relationship extraction], and [graph enrichment] techniques, while addressing challenges like [data heterogeneity] and [concept drift]. Use examples from [computer science] or [biomedical research] to illustrate your approach. Ensure the system supports [real-time updates] and [cross-domain interoperability].