An overview of most recent services and algorithms used by smart assistants - Google Now, Siri and Cortana
Introduction
<>
The brain of Cortana is all in Bing, so Bing powers Cortana. All the cloud services and all you can find on Bing, Cortana knows; for that reason she becomes smarter and smarter every day. Actually, when the user is doing speech recognition, when the user is talking to his phone, the focus is only on what Cortana says, but the user assumes that she will understand, because the speech recognition is that good.
How Cortana really works? Actually, when the user talks to his phone or desktop, Cortana is using a local speech recognizer and it is listening and it is actually streaming live; so whatever you are saying, …show more content…
The idea is simple — predict what you'll want or need to know before you know you need or want it, and serve it up in an easy-to-read format.The most formidable of the Siri and Cortana competitors is Google Now. Unlike Siri, which only runs on iOS, and Cortana, which only runs on Windows 10, Google Now runs on a variety of devices, and might work differently across different smartphones and operating systems. While Siri has a dedicated button, Google Now runs as a kind of intelligent layer under other applications on the phone, somehow like Cortana. In other words, even when you’re not saying “Okay, Google...” Google Now will still cue up information for …show more content…
The Dynamic Time Warping (DTW) distance measure is a technique that has long been known in speech recognition community. It allows a non-linear mapping of one signal to another by minimizing the distance between the two. A decade ago, DTW was introduced into Data Mining community as a utility for various tasks for time series problems including classification, clustering and anomaly detection[7]. Siri uses at first a discretization algorithm to turn your voice into digital data. These two steps are provided by Nuance Communication. Then it applies DTW to the digital voice to find phonemes. With the phonemes, it can figure out what words are being said. It also use grammar and other context clues to reduce the errors. After it turns voice to text, Siri uses an algorithm to detect keywords and to transform them into commands. After translating the text into commands, Siri checks if the requirement is a local problem or it is a global problem. For example, if the user wants to send a message, Siri detects the type of the demand as being local, searches through the local database and returns with an action, in this case opening the message section. If the demand is considered a global problem, the IOS assistant uses the Apple Server to detect what action should be taken or what information should be given. Unlike Google Now and Cortana, Siri does not have an important search engine behind, but the
Introduction
<>
The brain of Cortana is all in Bing, so Bing powers Cortana. All the cloud services and all you can find on Bing, Cortana knows; for that reason she becomes smarter and smarter every day. Actually, when the user is doing speech recognition, when the user is talking to his phone, the focus is only on what Cortana says, but the user assumes that she will understand, because the speech recognition is that good.
How Cortana really works? Actually, when the user talks to his phone or desktop, Cortana is using a local speech recognizer and it is listening and it is actually streaming live; so whatever you are saying, …show more content…
The idea is simple — predict what you'll want or need to know before you know you need or want it, and serve it up in an easy-to-read format.The most formidable of the Siri and Cortana competitors is Google Now. Unlike Siri, which only runs on iOS, and Cortana, which only runs on Windows 10, Google Now runs on a variety of devices, and might work differently across different smartphones and operating systems. While Siri has a dedicated button, Google Now runs as a kind of intelligent layer under other applications on the phone, somehow like Cortana. In other words, even when you’re not saying “Okay, Google...” Google Now will still cue up information for …show more content…
The Dynamic Time Warping (DTW) distance measure is a technique that has long been known in speech recognition community. It allows a non-linear mapping of one signal to another by minimizing the distance between the two. A decade ago, DTW was introduced into Data Mining community as a utility for various tasks for time series problems including classification, clustering and anomaly detection[7]. Siri uses at first a discretization algorithm to turn your voice into digital data. These two steps are provided by Nuance Communication. Then it applies DTW to the digital voice to find phonemes. With the phonemes, it can figure out what words are being said. It also use grammar and other context clues to reduce the errors. After it turns voice to text, Siri uses an algorithm to detect keywords and to transform them into commands. After translating the text into commands, Siri checks if the requirement is a local problem or it is a global problem. For example, if the user wants to send a message, Siri detects the type of the demand as being local, searches through the local database and returns with an action, in this case opening the message section. If the demand is considered a global problem, the IOS assistant uses the Apple Server to detect what action should be taken or what information should be given. Unlike Google Now and Cortana, Siri does not have an important search engine behind, but the