Telescope peering into the night sky

Microsoft Academic

Established: February 22, 2016

Editor’s note, May 4, 2021 – In a recent blog post (opens in new tab), it was announced the Microsoft Academic website and underlying APIs will be retired on Dec. 31, 2021.

 

Helping researchers stay on top of their game

Microsoft Academic (opens in new tab) is a project exploring how to assist human conducting scientific research by leveraging machine’s cognitive power in memory, computation, sensing, attention, and endurance. The research questions include:

  • Knowledge acquisition and reasoning: We deploy AI-powered machine readers to process all documents discovered by Bing crawler and extract scholarly entities and their relationships to form a knowledge base. To learn more and access our biweekly graph updates, please visit Microsoft Academic Graph (MAG) (opens in new tab) and its online documents (opens in new tab).
  • Semantic search and recommendation: The website, Microsoft Academic (MA) (opens in new tab), uses the same intelligence in the machine readers to infer query intent and retrieve most relevant knowledge in Microsoft Academic Graph (opens in new tab). Like an personal assistant, it can also recommend materials you might not know exist and alert you with the recent publication and late breaking news that you might find interesting.
  • Importance assessment and ranking: as needed in reasoning and inferences, the importance of each entity is estimated and quantified. We study reinforcement learning algorithms that can effectively predict community judgments on all entities, using future citations as the delayed reward function.

The most recent review of our work is published in this journal paper by Frontiers (opens in new tab) and this by QSS (opens in new tab).

Aside from the MAG data, a REST API powering the Microsoft Academic website is also available as a Cognitive Service Lab project (opens in new tab) for free-tier access only. There is no payment option to go beyond the throttling and quota limit. If you need faster responses and more calls than currently allowed, you are welcome to self-host the API. Please visit this online documentation site (opens in new tab) to see if self-host is right for you.

 

People

Portrait of Alvin Chen

Alvin Chen

Data Scientist II

Portrait of Darrin Eide

Darrin Eide

Principal Software Development Engineer, Microsoft Academic Services

Portrait of Rick Rogahn

Rick Rogahn

Principal Software Engineering Lead

Portrait of Iris Shen

Iris Shen

Principal Data Scientist