Nuance Adds Deep Learning to Dragon
Nuance Communications has begun leveraging its work with Nuance Deep Learning technology in the latest version of its suite of Dragon professional productivity solutions for PCs and Macs.
The company says that due to the Deep Learning technology, the new releases are 24 percent more accurate than previous versions.
The new Dragon releases include Dragon Professional Individual, version 15; Dragon Legal Individual, version 15; and Dragon Professional Individual for Mac, version 6. They each feature dictation, transcription, and customization capabilities and can sync vocabulary customizations and auto-texts with Dragon Anywhere to extend speech recognition to mobile devices running iOS or Android.
With deep learning technology at the core of its speech engine, Dragon can now learn individuals’ voice patterns and accents and adapt to acoustics in open-office or mobile environments. This latest Dragon release marks the first time deep learning capabilities can run embedded on a personal computer leveraging each user’s own speech data to improve language and acoustic models.
“We are all special in how we use language, from how we speak to the words we use,” says Rick Brown, senior director of product management at Nuance. “With our new Dragon Individual and Dragon Legal offerings, we are leveraging the uniqueness and individuality of each speaker to create even better user experiences and, ultimately, adapt to each user in a speaker-dependent way.”
Adding Deep Learning and neural net technologies allows Nuance’s Dragon offerings to adapt to the user’s active vocabulary by inspecting texts the user has created in the past, both by adding custom words to its active vocabulary and by learning the typical phrases and text patterns the user employs, Brown explains. During each session, a fast adaptation of its acoustic model takes place, capturing how words are pronounced based on just a few seconds of speech from the user.
“By doing this, the technology can also adapt to how a user’s voice sounds in the moment,” he adds. “Are they impacted by a cold, using a different microphone, or is there a change in environment?”
And it doesn’t end there. During the optional enrollment step, or later after the dictation session ends, Dragon does some more intense learning in an offline mode. “It continues to adapt models very well over time to a specific user’s speaking patterns,” Brown says. “Dragon uses deep neural networks end to end, both at the level of the language model—capturing the frequency of words and in which combinations they typically occur—and of the acoustic model, deciphering the smallest spoken units, or phonemes, of a language.”
The new Dragon also brings with it advanced transcription, customization, and an improved ability to work with portable touchscreen PCs.
Dragon Professional Individual, version 15 delivers additional usability and performance enhancements, including the following:
• optimization for popular touchscreen PCs;
• formatting and other recognition improvements;
• simplified setup and improved user interface;
• enhanced support for recorded audio transcription, with streamlined setup and improved profile management;
• expanded support, including full voice-driven editing, formatting, and command and control within popular desktop or Web applications for businesses; and
• contextual guidance tailored to user behavior.
In addition to leveraging the robust feature set of Dragon Professional Individual, Dragon Legal Individual, version 15 for the PC is designed specifically for the legal industry. A specialized legal vocabulary trained using more than 400 million words from legal documents enables legal professionals to create and format case files, contracts, or briefs all by voice.
Dragon Professional Individual for Mac, version 6 enables professionals to dictate reports, spreadsheets, email, and other documentation, and now offers a transcription batch mode to save time transcribing multiple audio files. Dragon Professional Individual for Mac supports full text control within the latest business applications. Dragon Professional Individual for Mac leverages Apple’s Accessibility API to support Apple Pages, Apple Keynote, Apple Numbers, Microsoft Outlook 2016, and Scrivener, among others.
Additionally, Dragon Professional Individual, Dragon Legal Individual, and Dragon Individual Professional for Mac can be synchronized with the Dragon Anywhere mobile app for Android and iOS devices, enabling a powerful cross-device experience so that professionals can use the same vocabulary customizations and auto-text commands wherever they use Dragon.
“This latest suite of professional productivity solutions brings with it some of the most advanced capabilities to drive documentation productivity with higher accuracy, speed, and efficiency,” said Peter Mahoney, senior vice president and general manager of Dragon, in a statement. “Combined with Dragon’s optimization for portable touchscreen PCs and the ability to sync with Dragon Anywhere, professionals are able to keep up with documentation demands from just about anywhere their business takes them.”
All versions of Dragon Professional with Nuance Deep Learning are available in English in the United States, United Kingdom, and Australia.