Skip to main content

KooKoo speech engine versus Google speech engine - Round 2

TLDR:

The data sets used for the testing in the experiments are the digit examples from the Google speech commands dataset.

Experiment 1:
Converted dataset to 8Khz to fit the telephony format.

Google API accuracy: 74.3%
Kookoo Zena accuracy: 92.4%


We felt the accuracy for Google dropped down because of the 8Khz format. So we did another experiment for Google speech API.

Experiment 2:
Original speech command dataset
Google API accuracy: 85.6%

The Long version:
At Ozonetel, we are constantly innovating and we are back again with a new speech model after the Yes/No model.

We all saw Google demo Duplex last week and it was a pretty cool demo. Just to put things in perspective, though the demo was cool, we think we are still some a ways away from actually having a system operate as smoothly as shown in the demo.
We come to the conclusion mainly based on our experiments with the Google speech API. Unless Google is using some model other than the one exposed publicly in the Google speech API, there is still a long way to go. Especially because the Google speech API is pretty bad with telephony data(8Khz).


Today, we are launching our second model, the single digit model, built in house, based on a proprietary AI algorithm. You can now design your KooKoo IVRs to say "Please say your order number", instead of "Please enter your order number".

In the internal tests our model's accuracy has been really good and we believe , soon, that this can become the standard interaction mechanism instead of DTMF.

We also did a side by side comparison of our model with Google's speech API. Before presenting the results below, some disclaimers:

1. Our model is a specific model for Digit. Google's model is more of a generic ASR. And in most cases specific models work better than generic models.
2. Google's ASR is in the cloud. Though our model is also in the cloud, since its co hosted with KooKoo, KooKoo responses will generally be faster.


The data sets used for the testing in the experiments are the digit examples from the Google speech commands dataset.

Experiment 1:
Converted dataset to 8Khz to fit the telephony format.

Google API accuracy: 74.3%
Kookoo Zena accuracy: 92.4%

We felt the accuracy for Google dropped down because of the 8Khz format. So we did another experiment for Google speech API.

Experiment 2:
Original speech command dataset
Google API accuracy: 85.6%

We also ran the test on multiple other private data sources. In all the cases KooKoo Zena out performed Google speech API.


Watch this space for more upcoming speech models.

Popular posts from this blog

First Post

In this blog, I will be talking about my experiences in trying to build a cloud telephony platform , KooKoo . Along the way I will also be talking about different design choices I made, good programming practices and the IVR domain in general. For technoratti: NNFJW8EW86C3

Google's approach to business communication

 Google has been making silent moves in the business communication space. Google has mostly lost the instant messaging wars. But it does not want to lose the business communication war. WhatsApp, Instagram, Twitter and Facebook have been making their own moves to enable businesses to reach their customers through their channels. Its all about who has control over the communication channels. Especially communication which leads to business. That's where the money is. Currently, Google is the king of search and most online transactions start with a Google search. FB, Amazon, Apple and others want to change that. They want the search to start on their properties. And they have started making the moves. WhatsApp business allows small businesses to conduct their transactions on WhatsApp. FB and Instagram have long supported small businesses to manage their business on their channels. Apple has also made some nice moves with Apple business chat. They have integrated a whole shopping expe

Integrating Arborjs with Angular to create a live calls dashboard

Arborjs  is a cool graph visualization library. Angular  is one of the best JavaScript frameworks and we have been using Angular in a lot of our front end development. When you handle millions of calls, proper visualization becomes very important. Without proper visualization, you can get lost in the mountains of data. So we spend a lot of time to come up with good visualizations to represent the data. Since we loved the cool way in which Arbor represented graph data, we could not wait to hook it up with Angular. Because of Angular's two way data binding, when you hook up Angularjs with Arbor.js you can get a dynamically updated visualization of graph data with cool animations. To give back to the community, we have put up the code online at Github . Basically we have created an Angularjs directive for Arborjs. Please feel free to fork the code and add extensions and use it for your own visualizations. The code is self explanatory with comments inline. Best way to get s