Google Voice Search Is Now Faster And More Reliable

Google Voice Search

The Google Speech Team have announced that they are improving their neural network acoustic models that will vastly improve Google Voice Search. To achieve this the team is using Connectionist Temporal Classification (CTC) and sequence discriminative training techniques, which are more accurate, especially in noisy environments, and they are blazingly fast.

This new technology developed by the Google Speech Team uses the entire sentence you speak instead of relying on individual word fragments to identify what you are saying. Using Recurrent Neural Networks (RNN) technology, Google is able to hear the sounds your voice makes in context.

Our improved acoustic models rely on Recurrent Neural Networks (RNN). RNNs have feedback loops in their topology, allowing them to model temporal dependencies: when the user speaks /u/ in the previous example, their articulatory apparatus is coming from a /j/ sound and from an /m/ sound before. Try saying it out loud – “museum” – it flows very naturally in one breath, and RNNs can capture that. The type of RNN used here is a Long Short-Term Memory (LSTM) RNN which, through memory cells and a sophisticated gating mechanism, memorizes information better than other RNNs. Adopting such models already improved the quality of our recognizer significantly.

To reduce computations, Google has also trained the models to take in audio in larger chunks while improving recognition in noisy places by adding artificial noise to the training data. The researchers said that to create the additional improvements, the speech team had to tweak the models to find an optimal balance between improved predictions and latency. These improvements give Google a faster and more accurate acoustic model that could be used on real voice traffic.

Google also mentions that they included artificial noise and reverb in their training data, which helps with voice recognition in noisy environments. In addition to being more accurate and quicker to respond, Google’s newer technology requires much lower computational resources. It is pretty technical, but if you want, you can read the entire blog here.

Google’s new acoustic models are already working in voice search for both Android and iOS, so feel free to try it out if you have not said “OK Google” in a while.

Source: Google Research Blog

Amarnath Natarajan Avatar

Help Us Grow

If you like this post, please share it with your friends.

You are free to copy and redistribute this article in any medium or format, as long as you keep the links in the article or provide a link back to this page.

Subscribe to Newsletter




Privacy Settings

Privacy & Cookie Overview

Our website uses cookies to provide you with the best user experience possible. These cookies are stored in your browser and perform essential functions such as recognizing you when you return to our website, as well as helping us to understand which sections of the website you find most useful and engaging.

To learn more, you can read our Privacy & Cookie Policy or reach out through our Contact form.

Strictly Necessary Cookies

Strictly Necessary Cookies must always be enabled to ensure the proper functioning of this website and to allow us to provide you with excellent service. These cookies are also essential for saving your cookie preferences.

Google Adsense

We use Google AdSense to keep this site free by displaying relevant ads. AdSense requires essential cookies that cannot be disabled, but you can manage other cookies. We respect your privacy and provide options to control non-essential cookies.

For more details on how Google handles your data, visit Google's Data Usage Policy. Please review our Privacy Policy for more information on how we protect your data.

AddToAny

We use AddToAny for social sharing. It doesnโ€™t store cookies, ensuring a privacy-friendly experience. AddToAny complies with GDPR and CCPA by default.

For more, see their Privacy Policy.

OneSignal

We use OneSignal to send notifications to users who opt in. OneSignal complies with GDPR and is certified under the EU-US and Swiss-US Privacy Shield frameworks.

For more, see their Privacy Policy.

3rd Party Cookies

This website utilizes third-party cookies, which can enhance your experience and support our ongoing efforts to improve our services.

Google Analytics

We use Google Analytics to collect anonymous data, such as visitor numbers and popular pages, to improve user experience and site performance. Keeping this cookie enabled helps us refine the site based on visitor activity.

For more information, see Googleโ€™s Privacy Policy.

Discover more from Prime Inspiration

Subscribe now to keep reading and get access to the full archive.

Continue reading