There are quite a few voice assistant gadgets or technologies such as Google Assistant, Amazon’s Alexa and Apple’s Siri. All these software tend to be useful in many ways. But the most important criteria to look at is to achieve a good accurate voice recognition by the gadgets or devices.
Click here to know more about Who is Best – Alexa, Siri or Google Assistant.
Only when the software is able to hear and understand us, the results would be fruitful. If the device interprets the words in a wrong way, then clearly it’s a downfall for both the user as well as the company.
This is exactly what Google’s CEO, Sundar Pichai realized. At the Google I/O 2017, he began his keynote by claiming how the Google’s voice recognition software has an error rate of less than 5 %. Ever since, it has been focusing a lot over this issue via machine learning and artificial intelligence. Google assistance is being pushed widely across many products since the past year, and hence to stand out among its competitors, it aimed at being as accurate as possible.
How improved is the Voice Command of Google Assistant?
Pichai noted that the voice assistant must be on par with humans in order to prove its accuracy. Humans can speak 150 words per minute on an average, whereas they can type just 40 words. This leads to the speech recognition technology to be a significantly dominant form of computing.
2017 Internet Trends report:-
According to the 2017 Internet Trends report, reported by Kleiner Perkins Caufield & Byers partner Mary Meeker, the following key points were noted –
- Google’s speech recognition is now on par with the human English language as it has a 95 % accuracy. This is same as the current threshold for the human accuracy.
- It’s accuracy has grown quickly and significantly to about 20% in the past 4 years (since 2013).
- There are few caveats, like how the error rate is calculated using real world search data which is more error prone than the typical human dialogue.
- Pichai also mentioned that, “error rates continue to improve even in noisy environments.”
- The improvements were mainly due to the efforts in Machine Learning Algorithms’ enhancement and Artificial Intelligence.
- A deep learning technique, called Neural Beamforming, allowed the company to release Google Home with just two microphones. But it has the same quality as of that having eight microphones.
- Voice queries are starting to overcome online questions, thus making it one of the most important internet trends this year. Voice has solved about 20 % of mobile queries.
- AI is also responsible for few recent features such as multi-user support. It recognizes and provides personalized assistance up to 6 users.
- Google Assistant can also schedule appointments, add reminders, and make hands-free calls using Google Home.
- Watch visual responses by using the assistant on TV via Chromecast. It also detects different accents much better than its competitors.
Google Assistant is one of the few AI voice-based digital assistants which can even understand contextual speech.
Its capability can be found in Google’s Home speaker, Android smartphones (Pixel and Galaxy S8), messaging app Allo, Android Wear 2.0, and also in the iPhone.
With all such remarkable updates, Google Assistant is surely helping its users move towards a smooth future.