课程: Azure AI for Developers: AI Design Foundations

免费学习该课程！

今天就开通帐号，24,700 门业界名师课程任您挑！

Connecting to structured data sources

Connecting to structured data sources - Azure AI Services教程

课程: Azure AI for Developers: AI Design Foundations

开始一个月试用 购买团队方案

Connecting to structured data sources

“

- [Instructor] So we've seen how to connect large language models to unstructured data, but what about structured data like spreadsheets, databases, and lakehouses? Well, that's not as easy as you might think. Large language models can certainly write code in query languages like SQL and KQL, but there are lots of mistakes they can make. They can make errors in the data schema. They can sometimes emit syntax errors in the code, and sometimes they just misunderstand the user's request. One way to solve for this is through the use of AI agents. Agents are a framework that sit above one or more models that allow you to do complex long running tasks and agents can take action on your behalf. They often involve chain of thought processes, which are more than a single large language model will do, and that can often allow them to recover from their mistakes. So agents are a good way to be able to query structured data. Later on in this course, we'll go a little bit more deeply into AI…

内容