Dataset Source Code
Structured Entity Extraction
Structured Entity Extraction and the Approximate Entity Set OverlaP (AESOP) metric are designed to appropriately assess model performance.
Research Tools: code, datasets, & models
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.
Dataset Source Code
AttentionEngine: A Custom Model Optimization Framework
AttentionEngine accelerates transformer attention variants by generating efficient custom kernels, enabling model designers to easily create new variants with our flexible API.
Dataset Source Code
TerraTrace: Spatio-Temporal Signatures for Land Use Analytics
Understanding land use over time is critical to tracking events related to climate change, like deforestation. However, satellite-based remote sensing tools which are used for monitoring struggle to differentiate vegetation types in farms and orchards…
Dataset Source Code
SeerAttention
SeerAttention is a learning-based method to enable block-level sparse attention for long-context LLM without using prefined static pattern or heuristic methods. It can be applied in Post-training or Fine-tuning stages. The Attention Gate units learn…
Tool
OmniParser V2
OmniParser is an advanced vision-based screen parsing module that converts user interface (UI) screenshots into structured elements, allowing agents to execute actions across various applications using visual data . By harnessing large vision-language model capabilities,…
Dataset Source Code
MICON (Molecular-Image Contrastive Learning)
This is the repository for paper “Causal integration of chemical structures in self-supervised learning improves representations of microscopy images for morphological profiling”. Learning effective representations of cells in microscopy images can fuel many applications. Here,…
Dataset Source Code
ProtNote: a multimodal method for protein-function annotation
ProtNote is a multimodal deep learning model that leverages free-form text to enable both supervised and zero-shot protein function prediction.
Dataset Source Code
AutoVerus
Automatically synthesize proof annotations that help Verus prove the correctness of Rust code.
Tool
FStar Data Set v2
This dataset is the Version 2.0 of the FStar Data Set. This dataset’s primary objective is to train and evaluate Proof-oriented Programming with AI (PoPAI, in short). Given a specification of a program and proof…