Research Tools: code, datasets, & models

Sent2vec maps a pair of short text strings (e.g., sentences or query-answer pairs) to a pair of feature vectors in a continuous, low-dimensional space where the semantic similarity between the text strings is computed as…

This download primarily contains a list of URLs with paired natural language descriptions and code, as well as a separate of those URLs into training, development, and test data. In addition, code is included to…

This is the source code for the Core Tabular command-line compiler, tc.exe.

Interactive Data Display for JavaScript (IDD for short) is a set of controls for adding interactive visualization of dynamic data to your application. It allows to create line graphs, bubble charts, heat maps and other…

Datasets and Python evaluation code used in the paper: S-MART: Novel Tree-based Structured Learning Algorithms Applied to Tweet Entity Linking, ACL 2015. Part of the data is based on the Making Sense of Microposts 2014…

This zip file contains CAD files and source code necessary to build and use an improved lens for the Oculus Rift HMD. The source code works with the Unity game engine to correct the lens…

MSR ECCLib is an efficient cryptographic library that provides functions for computing essential elliptic curve operations on a new set of high-security curves. All computations on secret data exhibit regular, constant-time execution, providing protection against…

Plato is a C++ open-source neural network library which supports the specification of a large range of graph types, several activation functions and training losses. The library supports backpropagation and truncated BPTT, especially useful for…

The R2 Probabilistic Programming Tool is a research project within the Programming Languages and Tools group at Microsoft Research on probabilistic programming. Our goal is to build a user friendly and scalable probabilistic programming system…

A collection of 12,696 Tweet Ids representing 4,232 three-step conversational snippets extracted from Twitter logs. Each row in the dataset represents a single context-message-response triple that has been evaluated by crowdsourced annotators as scoring an…