Voccent CEO Victor Brunko explains how the non-ML approach reduces barriers to learning a new language

(Photo: Victor Brunko, CEO of Voccent, explains how the non-ML approach reduces barriers to learning a new language)

Language is probably one of the most natural skills for human beings. Companies have spent years making their process of learning languages ​​as easy and straightforward as possible for people. In recent years, innovative technologies have revolutionized many areas of our lives and language acquisition is no exception. However, it turns out that it is the not-only ML (Machine Learning) approach that can really change the way we learn a language or acquire speaking skills.

How we learn languages ​​today

Humans are unique in that they possess complex linguistic abilities. We are able to communicate emotion and attitude “musically” using tone and inclination. Cultural and community attitudes are communicated through language. In addition, vocalization is combined with movement of the body and face to communicate information and emotions.

Or, even, why do we actually learn languages ​​and come to the limit – what do we have them for? It looks like we have languages ​​like any other living thing. Birds, cats, even viruses and bacteria have some kind of communication protocol. For us humans, the story goes much deeper. We not only “protocol” with each other using words. In fact, we communicate emotions and create all the flavor of the community by using the tone and inclination of a particular phrase. It could be a word or two that you’ve thought of on public channels, or something personal that you’ve decided to share with your neighbor. Yell at your dog? No worries, they understand our emotional background better than most primates, including … eh, humans.

Ok, convinced of the common truth – languages ​​are important. Thank you. How to learn them?

Well, you don’t have to “learn” to speak. They are, in fact, pre-integrated into most parts of the brain. Instead, we need to find an approach on how to activate these mechanisms. This is, as you would expect, language learning services, some dating as old as known mankind, exist. In the age of the internet and machine learning, the number of companies offering the best language learning experiences today is likely greater than the number of people learning these skills. Most of them are not even an app, it can be video channel or private voice chat with teacher in your best messenger.

How to break into this market? Who gets closer to what is needed instead of what is expected?

Acquiring vocal oral language is the key

Our approach is similar to learning music and engages the parts of the brain that manage tone and timbre (the unique tonal structure of an instrument). This allows learners to acquire stress and sentence structure as tonal units.

… and let’s go, Voccent is the company that takes on the ground from inside your skull. Well, they’re not going to drill and put on wires … yet, but it’s interesting. As it turns out, the biggest problem right now is that simple: repetition and neuro-feedback contribute to the maximum level of success one can achieve when learning new hearing experiences, and especially of spoken languages, relying on our brain’s oldest mechanism: Frequency Tracking Responses (FFR). Voccent deliberately does not distinguish between a myriad of language and dialect combinations at the software level, it is so blind in a good way to say this at audio inputs, that you can try to repeat after a musical instrument, like the piano or the guitar, or whatever else you just thought of. On this basis, Voccent CEO Victor Brunko says that internally within the team and with early adopters of the product, the agreement is to minimize the use of the words: “learn, teach, linear, lesson ‘in favor of the words:’ acquire, frame, non-linear, challenge ‘.

The technologies behind Voccent

Thus, to date, the startup is in a growth phase. Which means that all approaches and technologies are mapped to meet the global paradigm. Not without challenges of course. For example, you still have to make an effort. Listen to the samples. To browse Repeat-Compare-Improve-Repeat. Mitigate confidentiality requirements.
To achieve all this, a team of 8 people, including two co-founders, are busy building and supporting a variety of services. Critical services like speech processing and analysis are 100% local. The frequently asked question of young companies is: why not build on something that other big leaders in the speech industry provide as a service? Well, there is no short answer to that. Although the concise idea is that even the combination of different APIs that you will find today do not meet the requirements to realize this main idea: Voccent is a dialect-independent language acquisition service that allows people to put in practice their speaking / listening skills based on feedback. And the feedback can contain a number of variations: pronunciation, prosody, energy, pitch, rhythm, etc. Another issue is obvious: privacy and security. At Voccent, all audio files and non-public references are encrypted. Unless someone wants to share anything, all information is secure and hidden behind walls. Depending on the use case, for example, such as schools, government, or repeat customers, the privacy and security requirements are all different. The rule of thumb for the Voccent engineering team is pretty straightforward – even engineers with the highest level of infrastructure access won’t be able to listen to private audio. Sounds good as a minimum requirement.

People of the voice

How many people do you need to write 2 million lines of code? Well, it depends on focus and motivation. At Voccent Engineering, it took 8 people. With leaders preferring to warn of what not to do instead of just saying what to do, everyone on the team matters. The team doesn’t practice the standard approaches most companies do these days. In other words, it looks more like a science club than just another startup. It has drawbacks and advantages, but the bright side is the feeling people share – a prosperous future with the product doing something others don’t and something that can improve communication between humans.


Currently, the product is in the pre-production testing phase with a limited number of users. The Market Penetration Strategy team decided to simplify the user experience for streaming audio and video courses created by language teachers. This is achieved by adding the link to the language learning channel to the description on any other platform and getting customers to practice the target language on the Voccent platform. The first results and feedback show that it is more effective both for content creators (* tubers, * grammers, * tokers) and for people who acquire new language skills. Visit https://voccent.blog/ to learn more about Voccent.

2021 TECHTIMES.com All rights reserved. Do not reproduce without permission.

Key words:

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *