2.7 Data Science Agent

Data Science Agent is an intelligent built-in conversational chatbot integrated with Oracle Machine Learning UI included in your Oracle Autonomous AI Database subscription. You must provide the LLM, whether from a third-party AI provider, OCI GenAI Service, or one you privately host. You can run many common data science workflows, including discovery, exploration, preparation, model training, evaluation, and supported scoring workflows, using natural language.

As a human-in-the-loop chatbot, Data Science Agent requires your intervention and approval in the following areas:

• Object association: Associating database objects with a conversation improves precision and efficiency. Objects must be associated or approved before the agent creates views that use them; other supported operations may still use accessible objects based on the user's database privileges. See Add Objects to Data Science Agent Conversation.
• Conversation management: The scope of Data Science Agent is limited to the active conversation context only. You must create, manage, and delete your conversations. See Manage Data Science Conversations.
• Profile selection: You must select an AI profile, and therefore, the LLM that powers the agent. You can also switch profiles mid-conversation. See Manage AI Profile and Service Level of Data Science Conversation.
• Approval for object view: During any conversation, Data Science Agent compares the object_list with the objects already associated with the active conversation. If required, the agent proposes any table or view outside the user-approved set. It stops the conversation and explicitly seeks your approval to associate those missing objects before continuing. If you approve associating the missing objects, the association is recorded and the agent continues the conversation. Else, the operation is canceled and no view is created.