Big Advancements in Artificial Intelligence and Speech Recognition

Small Steps for Microsoft and Others Will Mean Big Steps for User Experience


In today’s world, information has to be precise and delivered in time to provide the best results. As a result, the common issues with chatbots and their failures is becoming quite cumbersome for consumers as more businesses depend on Artificial Intelligence systems as their front line customer service. To tackle this issue, Microsoft has been working with different ways of making machine voice recognition and interpretation more effective.

Advancements in AI

According to a paper published recently, researchers working for Microsoft have managed to attain a relatively low error rate for speech-recognition software, standing at just 6.3%. The company is hopeful that they will be able to personalize and refine their AI assistant, Cortana, as well as other features such as Skype Translator to improve user experience (ux) with the products.

According to Microsoft’s speech scientist, Xuedong Huang, their new conversational speech-recognition system is currently recognized as the best in the industry having achieved the lowest error rate thusfar. Another tech giant to declare its stats was IBM, announcing an error rate of 6.6%, which was an improvement from its earlier standing of 6.9% and 8% last year. According to a blog post put up by Microsoft, the lowest rate of any system available two decades ago was more than 43%, which clearly shows you how far chatbots have come in a short time.

A few other technologies that have come about in the field of AI include biologically inspired systems such as neural networks, a deep training technique known as deep learning, and the implementation of graphic processing units or GPUs, to process algorithms. Over the past couple of years, deep learning and neural networks have allowed AI researchers to create and train systems related to advanced speech recognition, natural language processing, and image recognition. In addition, over the last year, Microsoft has also managed to create image recognition software that is capable of outperforming humans and may, relatively soon, allow users to be recognized universally on all devices they use.

Even though it was initially intended for computer graphics, GPUs are now commonly used to process complex algorithms. According to Microsoft, Cortana is capable of processing up to 10 times more data using GPUs than it could with previous methods.

AI and the aspect of user experience

When it comes to artificial intelligence development, researchers tend to model machines to resemble human-like characteristics, primarily speech, sight, and hearing. Despite the fact that Microsoft’s milestone is just a mere 0.3% below IBM’s, incremental steps such as these are essential when it comes to web engineering, UX design, and IoT. In addition, they bring machines steps closer to achieving near-human intelligence.

With such steady advancements, it means that constantly repeating your questions to a chatbot will become a thing of the past. With clear information, eCommerce websites stand a chance to generate more revenue and companies who rely on automated systems for customer service will vastly improve their customer’s experience dealing with bots and other types of AI.

Posted October 4, 2016 by & filed under IoT, News.