Discover an index of datasets, SDKs, APIs and open-source tools developed by Microsoft researchers and shared with the global academic community below. These experimental technologies—available through Azure AI Foundry Labs (opens in new tab)—offer a glimpse into the future of AI innovation.
Bing Artificial Search Sessions(BASS) is a collection of 18m Artificial Search session that were created by taking real conversational Search Sessions and mapping them to publicly available queries using vector space embeddings.
MS MARCO is a collection of datasets focused on deep learning in search. The first dataset was a question answering dataset featuring 100,000 real Bing questions and a human generated answer. Since then, we released…
Access Publication Publication Publication Publication Publication
A python package with a reinforcement learning algorithm that decodes latent states from rich observations.
The Bosque programming language is an experiment in regularized design for a machine-assisted rapid and reliable software development.
The Quantum Development Kit includes Q#, a brand-new quantum-focused programming language with native type, operators, and other abstraction, advanced code optimization in a simulated environment and a collection of ready-to-use open source libraries and samples…
The NL2API dataset includes the web APIs call from the Microsoft Graph API suite, which are respectively used to search a user’s emails and calendar events. Each data points include the API call, its canonical…
Visual Navigation via Language-based Assistance
A set of Databricks and Azure notebook scripts that compute urban heat island effects for all major cities in the United States. Daily indices for each city are computed from January 1, 2008 to present…
This PyTorch package implements the Multi-Task Deep Neural Networks (MT-DNN) for Natural Language Understanding.
We introduce a new corpus of descriptions of Xbox avatars created by actual gamers. Each avatar is specified by 19 attributes, including clothing and body type, allowing for more than 10^20 possibilities. Using Amazon Mechanical…
