Microsoft Research: Advancing Data, Knowledge, and Intelligence

Data, Knowledge, and Intelligence at Microsoft Research

Microsoft Research's Data, Knowledge, and Intelligence (DKI) group is dedicated to democratizing data intelligence, empowering individuals and organizations to derive insights, share knowledge, and translate data into actionable outcomes. Their research spans various forms of data, focusing on three core themes: understanding data, generating data, and interacting with data to create unparalleled user experiences.

Core Research Areas:

Data Analytics Research: This area focuses on understanding data, modeling analytical processes, and developing techniques for automatically generating insights and reports. Key technologies include representation learning, schematization of data artifacts (like tables and questionnaires), insights mining, forecasting, and causal inference. The underlying technical pillars involve machine learning, multi-dimensional data mining, explainable AI, and graph models. These technologies have been integrated into Microsoft products such as Office (Excel, Forms, Word), Power BI, Dynamics, and Bing Search.
Visualization and HCI: This research area explores information visualization, visual analytics, and human-computer interaction. By leveraging machine learning and AI, the group develops novel technologies, user interactions, and systems to lower the barriers for users in data analysis and communication. Specific focuses include visualization and infographic design, data-driven storytelling, visual analytics systems, new user interfaces for data exploration, and data wrangling.
Natural Language Understanding for Data Science: The mission here is to advance natural language understanding (NLU) technologies for intelligent data science. This is crucial because a significant portion of real-world data is unstructured natural language, requiring NLU for analysis. Additionally, natural language serves as an intuitive interface for data analysis tools, enabling common users to explore data through natural language instructions. Research topics include semantic parsing, semantic role understanding, entity recognition, and dialog systems. Fundamental machine learning problems, such as compositional generalization in DNN models and reinforcement learning sample efficiency, are also addressed to support NLU research. These NLU technologies have been implemented in features like Excel Ideas, Power BI Q&A, Microsoft Bot Framework, Azure Text Analytics, and more.
Knowledge Computing: The Knowledge Computing Group aims to build machines that effectively utilize knowledge to empower individuals. Their primary focus areas are natural language processing (NLP), information extraction (IE), table interpretation (TI), and knowledge representation & reasoning (KRR). NLP analyzes, understands, and generates language for human-machine communication. IE extracts structured data from natural language texts. TI identifies column types, cell entities, and relations in tables for question answering. KRR provides the symbolic representation and reasoning capabilities for NLP, IE, and TI. Technologies from this group have been integrated into Office 365, Azure Cognitive Services, Bing, Microsoft Forms Design Intelligence, PowerPoint Designer, Excel Data Types AutoDetect, Azure Text Analytics, Microsoft Video Indexer, and Microsoft Recognizers Text.
Software Analytics: This area focuses on leveraging the vast data generated throughout the software lifecycle (source code, specifications, bug reports, logs, user feedback) to improve software quality and development productivity. By employing technologies like pattern recognition, machine learning, data mining, and large-scale data computing, software analytics enables practitioners to explore and analyze data for actionable insights. The group at MSR Asia specifically aims to advance this field and apply its technologies within Microsoft and the broader software industry. In recent years, research has expanded to Cloud Intelligence, using AI/ML to manage and operate complex cloud services, focusing on AI for Systems/Infrastructure, AI for Customers, and AI for DevOps. These technologies have been transferred to services like Azure, Office 365, and Bing.

Engagement and Community:

The DKI group actively engages with the research community and the public through various channels:

Social Media: They maintain a presence on X (formerly Twitter), Facebook, LinkedIn, YouTube, and Instagram to share updates and engage with their audience.
Community Interaction: They encourage following their social media channels and subscribing to their RSS feed for the latest news and research.
Sharing: Options are provided to share content via X, Facebook, LinkedIn, and Reddit.

Key Products and Technologies:

Many of the research outcomes from the DKI group have been successfully integrated into Microsoft's products and services, including:

Excel (Ideas, Data Types AutoDetect)
Power BI (Q&A)
Dynamics
Bing Search
Microsoft Forms (Design Intelligence)
PowerPoint Designer
Azure Cognitive Services (Text Analytics)
Microsoft Bot Framework
Microsoft Video Indexer
Microsoft Recognizers Text (open source)

The research in Data, Knowledge, and Intelligence at Microsoft Research is at the forefront of leveraging data to drive innovation and empower users across a wide range of applications and industries.