As a native Hebrew speaker, Yael Karov said she’d often search the web for English phrases to make sure they were correct before including them in emails and other documents. Via the web, for example, she’d learn that “on July” should be “in July” or that “3 millions” should be “3 million.”
But given her background in statistical models and natural language processing, she knew that there should be a less tedious way to perfect her writing. Now, with her startup Ginger Software, she’s putting algorithms to work to help other English language learners correct their writing and try to improve it.
Since launching in 2008, the company, which has raised $20 million in venture funding, says it’s mapped one trillion English sentences on the web. Its “Proofreader” and writing “Coach” apps have been downloaded three million times (the company declined to share how that breaks down in terms of paid vs. free apps but offers the Rephraser and Proofreading apps for free and charges users $59 for the company’s premium package, which includes those two apps, the coaching app and a text to speech product.)
This week, Ginger rolled out its latest tool – a “Rephraser” app that helps users fill in missing words in sentences and suggests alternate phrases to help them vary their writing or find different ways to express themselves.
The new app integrates with web and desktop tools, including the Chrome and Safari browsers,Microsoft Word and other desktop applications. When a user types in a phrase, it presents a list of the most frequent and relevant alternates generated by a statistical algorithm that analyzes the online usage of English phrases.
“The idea is that the correct way of usage is much more frequent than the incorrect way,” Karov said.
For example, if you type “thanks for the meeting,” it might offer up “thanks for taking the time to meet” or “thanks so much for meeting with me.”
For English language learners working with clients or applying for jobs where English is the primary language, Ginger’s app could be a valuable tool. But it also seems as though the app could end up simply being a crutch — instead of processing language mechanics, users could just lean on technology to serve up the right phrases. (Sure, that could be a knock against spellcheck and proofreading apps of all kinds, but providing entire alternate phrases seems to require even less work on the part of the writer.)
To address that concern, Karov said, the company last month released its coaching app (Ginger Coach). The product analyzes a user’s written language over time to identify mistakes and diagnose her biggest areas of weakness. From there, it provides her with customized lessons intended to improve those skills.
While Ginger Software is taking on the English as a Second Language market in an interesting way, it’s one of several companies bringing natural language processing technology to consumers. Apple’s Siri may be the most high-profile example of natural language processing at work. But Nuance Communications offers a range of products that bring voice and speech technology into the home and office. And startup Duolingo uses natural language processing to help speakers of several different languages learn new languages while translating the web.
Image by sunabesyou via Shutterstock.