Recherche
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.
The HoloLens Navigation for Robots project is an open source repository to be included as a release in the Applied Robotics Research Library. It is a set of sample code to build software capable of…
The purpose of this project is to release a set of clinically relevant labels for an existing, publicly available dataset (https://fastmri.org). With permission from the fastMRI team at NYU, we have engaged third party labelers…
Adversarial GLUE Benchmark (AdvGLUE) is a comprehensive robustness evaluation benchmark that focuses on the adversarial robustness evaluation of language models. It covers five natural language understanding tasks from the famous GLUE tasks and is an…
The ORBIT dataset is a collection of videos of objects in clean and cluttered scenes recorded by people who are blind/low-vision on a mobile phone. The dataset is presented with a teachable object recognition benchmark…
The Tip of the Tongue (ToT) dataset is from the paper Tip of the Tongue Known-Item Retrieval: A Case Study in Movie Identification (opens in new tab). It is comprised of 758 question/answer pairs scraped…
Microsoft sponsored and co-organized Indoor Location Competition 2.0 in 2021. 1446 contestants from more than 60 countries making up 1170 teams participated in this unique global event. In this competition, a first-of-its-kind large-scale indoor location…
This is a codebase for our recently released paper “Focal Self-attention for Local-Global Interactions in Vision Transformers”. It developed a new sparse self-attention mechanism called focal self-attention towards more effective and efficient vision transformers. The…
The subseasonal_data package provides an API for loading and manipulating the SubseasonalClimateUSA dataset developed for training and benchmarking subseasonal forecasting models. Here, subseasonal refers to climate and weather forecasts made 2-6 weeks in advance. See…
Data augmentation is proven to be effective in many NLU tasks, especially for those suffering from data scarcity. In this paper, we present a powerful and easy to deploy text augmentation framework, Data Boost, which…
Microsoft KaggleDBQA is a cross-domain and complex evaluation dataset of real Web databases, with domain-specific data types, original formatting, and unrestricted questions. It also provides database documentation, which contain rich in-domain knowledge. The nature of…