An AI model must follow a strict process for data collection and enrichment. The model must not generate content from memory when actual text is available. Each step in the data pipeline must be completed in order.
By Pawan Nayar · from the LLOS AI Learning record
An AI model must follow a strict process for data collection and enrichment.
The model must not generate content from memory when actual text is available. Each step in the data pipeline must be completed in order.
A developer can build a four-step process. This process ensures accurate data collection. It prevents the AI model from creating information. The goal is to gather a large vocabulary from sacred texts. This vocabulary will cover many religious traditions.
A developer needs to create a list of sacred words. The developer uses the four-array architecture. The developer first lists religions and their texts. Then, the developer fetches the actual text for each book. Next, the developer extracts unique words and phrases. Finally, the developer uses an AI model to add meanings to these words. This ensures the AI model does not invent words or meanings.
Resolved from this lesson's own record — real definitions, not invented for this page.
More from the same record, pulled by shared topic — not written new for this page.
A developer must check all slogans in the AI model's training data. This check happens in two steps.
The language in an AI model's user interface defines its personality. Using human-like words makes the experience more engaging.
An AI model can understand and respond to questions that are not common. These questions are often called 'long tail' queries.
The AI model first creates a list of unique words. This list comes from an earlier check.
An AI model must follow a strict process for data collection and enrichment. The model must not generate content from memory when actual text is available. Each step in the data pipeline must be completed in order.