1) The document discusses developing cognitive agents to improve deep question answering and discovery by coupling two platforms: the Berkeley Data Analytics Stack (BDAS) for big data analysis and SolrSherlock for literature-based discovery. 2) It describes how these agents could harvest and represent patterns, contexts, and relations from literature to discover new processes and connections between concepts. 3) The goal is to augment existing methods by allowing hypothesis formation and evidence gathering across both structured data and unstructured literature at scale.