Microsoft Research has a group in Redmond and another in Beijing working together to improve spoken language technologies. Our main goal is to build applications that make computers available everywhere, and work with Speech Platforms Group product group to make this vision a reality. We are interested not only in creating state-of-the-art spoken language components, but also in how these disparate components can come together with other modes of human-computer interaction to form a unified, consistent computing environment. We are pursuing several projects to help us reach our vision of a fully speech-enabled computer.

In Redmond we are working on multimodal user interfaces, and that helps us discover real problems that we need to solve to make speech recognition more useful. The first such application we built was MiPad, a personal information manager used with speech and pen. We are also working on several other areas, including noise robustness, acoustic modeling, language modeling, grammar induction, SALT (Speech Enabled Language Tags) and personalization.

Flash overview of speech recognition at MSR (click in "Microsoft Research" and "Speech technology"). We have a few videos that illustrate our technology and a list of publications for more detail.

For information about the source code that we have made available to the research community check out our downloads page.

Primary Contact: Alex Acero
Acero, Alex	Deng, Li	Droppo, Jasha	Li, Xiao
Nguyen, Patrick	Seltzer, Mike	Tashev, Ivan	Wang, Ye-Yi
Yu, Dong	Zweig, Geoffrey
Affiliate Members
Bohus, Dan	Brill, Eric	Chang, Eric	Gunawardana, Asela
He, Xiaodong	Hon, Hsiao-Wuen	Huang, Xuedong	Malvar, Rico
Paek, Tim	Zhang, Zhengyou

XD Huang	xdh	Redmond
Kuansan Wang	kuansanw	Redmond
Hsiao-Wuen Hon	hon	Redmond
Fil Alleva	fil	Redmond
Li Jiang	lij	Redmond
Mike Plumpe	mplumpe	Redmond

• Noise Robustness: How do we make the system work when background noise is present?
• Microphone Arrays: Noise reduction through microphone arrays.
• Dereverberation: Combatting reverberation.
• Acoustic Modeling: How do we model phones and acoustic variations?
• Language Modeling: How likely are words?
• Automatic Grammar Induction: How do create grammars to ease the development of spoken language systems?
• SALT (Speech Enabled Language Tags): A markup language for the multimodal web
• Multimodal Conversational User Interface
• Personalized Language Model for improved accuracy

In the past, the speech technology group has worked on other projects, which have been successfully completed, and are either in shipping products (through the Speech Platforms product team), or have moved to the product development stage. These include:

• (MiPad) Multimodal Interactive Pad. Our first multimodal prototype.
• (Whisper) Speech Recognition Our previous dictation-oriented speech recognition project is a state-of-the-art general-purpose speech recognizer.
• (Whistler) Speech Synthesis (Text-to-Speech). We have produced a speech synthesizer so that your computer can talk to you.
• (WhisperID) Speaker Identification. Who is doing the talking?
• Speech Application Programming Interface (SAPI) Development Toolkit. The Whisper speech recognizer can be used by developers to produce applications using speech recognition

Click for full size