Microsoft Academic Search is Microsoft's version of a library database. It contains records for nearly 200 million articles (as of 2/2018). However, it's the way that MA treats those articles that is different from a traditional database. The articles have been analyzed for semantic content, not just indexed for keywords, so related concepts will automatically be searched together. This means the user is not as vulnerable to quirks in the authors' phrasing. MA also looks at the relationships between articles, based on references, citations, and authors, as well as "semantic similarity."
These videos will give you some ideas on how MA can be used. But the best way to explore this is to try it out. It's not quite like searching a regular database.
Transcript coming soon
Transcript coming soon