June 14, 2024

What Will Language Learning Look Like In The Future? Voccent CEO Victor Brunko Explains

Language learning technology has been a subject of research for over forty years now. Computers, smartphones, artificial intelligence – these are some examples of breakthroughs that totally changed the process of language learning. Scientists and IT experts believe that machine learning (ML) will play a major role in both learning and teaching techniques in the near future. However, is this the only approach? According to the concept by the Voccent company and it’s founder – Victor Brunko – no it is not. A not-only-ML approach can be greatly efficient as well.
The Role Of ML In Modern Language Learning Apps

ML is penetrating our daily lives. It is already an integral part of financial technologies, online marketing, gaming and e-commerce. A few years back, these technologies started to be implemented in language learning education services. The learning process has become highly modernized, allowing us to take a giant leap forward. Machine intelligence has improved the quality of translations and incorporated machine translation into foreign language learning. The latest ML innovations have made it possible to adapt the learning process to suit students’ needs based on their individual data. Nowadays, we have arguably become too obsessed with Machine Learning and Artificial Intelligence as a whole.

Victor Brunko, CEO of Voccent, decided to take an alternative not-only-ML approach for the project his team is working on, and so far, it seems to be very promising. They admit that there are, of course, some benefits to using AI for language learning, as it can help to create dialogues and extract lots of interesting features from an audio-context, lets say – an accent. However, ML has proven to be insufficient on its own, energy-consuming and follows the next paradigm: “created to be trained”. Amazingly, the solution that the Voccent team found to be most effective is a pure Digital-Signal Processing (DSP) combined with mathematics and a few subtle ideas from neuroscience.
Voccent Learning Idea

The overall idea of learning is pretty simple for the living beings: we all have our brains, those have two types of data allowed from the very beginning: genetic based and acquired. When we learn, new experiences cause changes in brain structure. The more frequently the input occurs – the stronger the connections between neurons. When we experience positive feelings about newly acquired skills, established connections become deeper and stronger. This is known as reward reinforcement learning.

Where does this lead us? We need to have feedback from the outer world. That’s where modern learning applications should come into the game… until they are not. At least most of them aren’t.

Let’s take as an example the most complex in terms of abstraction level one of the disciplines people learn – Language Learning. What do we mean by that? Reading? Writing? Speaking? Listening? A combination maybe? Recall, when you first learned the simplest concept from math for example? A 2+2=4 is a fair statement. But do we really calculate it all the time? Or do we remember by heart? Well, research shows that well-trained subjects consume less energy and don’t spend time adding 2+2 all the time – they simply recall answers.

Languages are arguably more complex in terms of potential combinations we can retrieve. Plus it may have four different modes of retrieval – reading/writing/listening/speaking. And the brain handles this just fine… when trained.

Having said all that, what do we really want from the Language Learning applications? Per Voccent founders Victor Brunko and Lev Kaufman there are many things we first need to un-focus from. Either we handle auditory or spelling and grammar. People can handle both for sure, but a laser-focused brain can make miracles. And Voccent is all about active speaking and listening. It’s all about sounds. Even brain music… but it’s for a later conversation.

To summarize a rather long explanation – if one learns how to Speak and Listen, it makes sense to focus on these practices and train your brain using positive feedback to match the sounds that trained (native) speakers use in conversation (imitative learning).

This is why the founders prefer the term “Speaking skills acquisition” over “Language learning” when they describe Voccent platform. You’re not going to talk with Machines (at least this decade for sure), we’re going to communicate with humans. And humans are all different.
People Of Voccent

Ok, thanks you’d say. Can you upload data to my brain? Yes, we’re almost on it. But first, let’s talk about humans and not only who made this project happen. As it goes from the ground up, the founders of the project are with long-lasting experiences in the startup business as well as in big players. Victor Brunko, a Voccent founder and CEO, has a long-lasting story of creating companies starting back in Russia and now in the United States. Penetrated with the good vibes from Tesla Palo Alto, California, and other companies. Lev Kaufman has an even deeper knowledge base with most of his experience coming from telecom companies in the discipline of customer support and digital-signal engineering. Many others made this project happen and keep it running, we’ll talk about it in the next conversations, but to keep from the main discipline, it’s worth mentioning Ana Iriarte Ruiz (PhD. In Telecommunication Engineering and Computer Science by Autonomous University of Madrid (2012), Associate professor in CEU San Pablo University (Madrid) from 2008 to 2016). That’s where the magic and sorcery are taking roots from. Overall Voccent Corp is a team of around 10 people and growing. More about them on the https://voccent.blog/ link.
So, What Are Their Plans For The Future?

The Voccent platform is able to differentiate between the multiple levels of language and communication complexity. It can recognize a variety of voice nuances, like accents or even emotions behind something, said. The prosody analysis, which is definitely Voccent’s cool feature, makes it possible to pinpoint emotional context, tone of voice, pauses, and exclamations. This will be a great competitive advantage that is sure to open many lucrative doors for Voccent – not only in language learning but also in technology, health care, HR, and perhaps in more serious industries, such as the government or military. We are excited to see what the future will bring!

See more from Benzinga

© 2021 Benzinga.com. Benzinga does not provide investment advice. All rights reserved.