Global Tech Firms Are Using Machine Learning To Translate The Vernacular

Artificial intelligence, Neural network language

The early 2000’s notion that Internet will eliminate the need for translators – or even the need to learn foreign languages at all – seems to get real in the next few years. 

Going half a century back, on January 7th, 1954 the world witnessed the most famous machine translation(MT) demonstration, today is known as Georgetown-IBM experiment, the event showcased the power of machine translation- an IBM 701 translated more than 60 sentences from Russian to English, without any human intervention. The experiment was much publicized, leading to a widespread belief that MT would completely replace human translator with a few short years.

In countries like India, with its large linguistic diversity, it is quite essential in the age of Internet that the technology of machine translation(MT) for Indic-language scripts needs to be further explored, both as a matter of convenience for native users as well as a means to expand reach and scope of business.
India being the second largest Internet market, much of the population in the country is yet to come online. One of the reasons that continue to thwart their efforts, the world wide web is the huge language barrier. Most Indians are not proficient with English, and much of the web and apps made available to them is not available in their native vernacular languages.

This is the sad reality of emerging markets where the literacy rate is often poor and local infrastructure is not conducive to bridge that gap. It is also a major challenge for many firms, including Silicon Valley companies that are hunting for their next billion users in places like India. But few among them are also look at it as an opportunity to showcase their engineering talent.

Global internet giants like Facebook, Amazon, Google and Microsoft are developing advanced Artificial Intelligence(AI) program to cater to this demand. Facebook allows users to post content in more than 9 Indian languages while Amazon furnishes documentation and online support in Hindi for its Indian sellers. Google, the world’s largest search engine, supports a number of Indian languages, Microsoft provides e-mails in 15 Indian languages across its apps and services.

Robot, Artificial Intelligence,

Today, Artificial intelligence (AI) has crept into numerous aspects of our lives, thanks to improvements in the field of machine learning, where computers ostensibly program themselves. This drive towards computer self-learning has led to major breakthroughs in our day-to-day interactions with machines, most notably the rise of digital home assistants such as Amazon Echo, and the recently launched Google Lens, which recognises objects based on visual cues from phone’s camera.

For instance, Google has developed Google Neural Machine Translation (GNMT) for translation from English to nine Indian languages. GNMT is better equipped to understand contextual differences and renders human-like speaking ability to the machines. Last year, Microsoft also announced the integration of Artificial Intelligence (AI) and Deep Neural Networks (DNN) to improve real-time language translation for Hindi, Bengali and Tamil.

As Natural Language Processing and Understanding (NLP/NLU) technologies further developed from text to now voice and speech synthesis, and as voice-based applications become commonplace, language localization needs and aiding AI & ML technologies will also evolve to speech-to-Text, NLU engines.

We have already started witnessing the early adopters. These include e-commerce giant’s Amazon’s Alexa being prepped up for Hindi, and the Google map Hindi voice guide.

Once neural network-based speech recognition engines turns mature for Indian vernaculars , they will find wide acceptance across applications including voice-based searching from web-based applications, query search on platforms of e-commerce, health, education, information media, entertainment, etc., video/audio transcription for subtitling and captioning, analysing audio for semantics and categorization, voice-based assistants.

More