AI and Language Learning… Does It Still Make Sense to Learn a Foreign Language?

hey friends welcome to the channel should you still learn a language nowadays soon Ai and machine translation will make that obsolete right hasta la vista baby these are the topics we'll talk about today no skates my name is Dustin I'm an ID portrait manager from Germany on this channel we create content that helps us we are productive to learn new things especially languages and to live a happy hour of filling life today we'll look into if it still makes sense to learn a foreign language since it's a huge time commitment and we can do many other things so it just makes sense to ask ourselves why we are doing it and if it's worth it therefore we first have a look at the history of machine translation since it has been invented further we will review the available gadgets and tools of today which can be used by regular folk like you and me and we will have a look at what's to come and last but not least we will answer the question if it still makes sense to learn for our languages okay let's dive into today's topic but first of all a nice sip of coffee before we get into the history of machine translation let's have a look at what the main challenges for them are and the main thing is the context so how a text is written and how it should be perceived and to distinguish the nuances the language is very complex it's really hard to distinguish in the context since the context is not just one thing but there are many factors which influence it first who is the recipient what cultural background does the recipient have and also the writer where does everybody come from which age do they have which occupation and further what the text is about and its underlying implications and subtopics so what is written between the lines and also when you take into account Concepts like sarcasm or irony or humor so now let's have a look into the history of machine translation it all started in the 50s with rule-based translations this means that dictionaries and grammars are used to make translation but of course this approach has its limitations and this concept really struggles with the context then the next Evolution step has been statistical machine translation which started in the 90s and then evolved really in the 2000s this means that you take bilingual text and then generate statistical models based on that which you then use to do your translations and this approach also struggles to take into account the complexity of the context but it's already more advanced than the former approach and definitely gives more accurate translations in our state of the art is neural machine translation and let me quickly quote Wikipedia here neural machine translation nmt is an approach to machine translation that uses an artificial neural network to predict the likelihood of a sequence of words typically modeling entire sentences in a single integrated model so we actually model an artificial neural network which is yeah inspired by the structure of our brain and this approach has been largely popularized by Google in 2016 when they switched Google Translate to such a model it's not just more accurate than the other approaches but it also takes way less memory and the whole technology is also evolving further so we have for example highly specialized domain translations which means that there are large quantities of data being gathered in a specific domain so that the translations can be really accurate and this is especially interesting for large companies like insurance companies or whatever to automate and to save money then we have also adaptive machine translation and this is really interesting because it's like the name already says adapts the translation while it's translating a text or a dialogue so it will take the stuff into account that it has already translated for the upcoming sentence so this really yeah it's also taking into account context there are earbuds for translations and those resemble airpods if you will and there are also specialized companies like for example time cattle they do exactly that and yeah you've got two of them you put one in your ear the other one into the year of the what the person you are talking to and then you start the app and you were just talk in your mother tongue and yeah the app would transcribe it and translate it and then the recipient would hear the translation in its earpiece and you could then or she could then reply and the same spear would happen again you would get the translation and that yeah immediately into your ear and this works quite well I've been really impressed how well actually I've seen a video about that by a big bong and I'll also link it here and yeah I have a further look it's really interesting and there are also other companies like for example Google who offers something like that the Google pixel Bots I think they are called and yeah those should also work in theory but I think time cattle is way more specialized and probably a bit more accurative than online meetings with real-time translation so of course that's not speech translation in that case or in that sense but it's transcribing what you are saying and then translating it so it's like having a real-time conversation but with subtitles like if you're watching a series on Netflix or something and I found that pretty impressive I haven't tried it out yet but I've stumbled a bit through the internet and I've read some stuff about that and it looks quite interesting so I could really imagine that this is the way going forward then there are also online translation tools like I've already mentioned Google translate which has a huge variety of available languages and it's not bad but there's one translator I like even more that is deep L this translator also uses state of the art technology and it has way fewer languages than Google for example but especially in the European area it has most of the the available languages and the translations are more accurate then there are also transcription services so where you would send to an audio file and they would transcribe it and sometimes even translate it still the majority of them operates with human translators but there are more to come with machine transcriptions and also machine translations and this will further lower the cost and can be really interesting for example there's a company called trint which are doing that and which support many languages and I'm pretty excited about that because it will further lower the cost and will enable me to also get transcriptions for text that I could send them and which I could then use to learn a foreign language and then of course there are more and more languages to be translated also smaller languages and there's a project called a no language left behind by meta and they try to do it exactly that so together more material and more resources on smaller languages and to also make the algorithms work with fewer resources so first of all stuff will evolve further so everything I've just shown you will get better there will be more language if it's available for translations context will be more in the focus also there will be more accuracy with the translations transcriptions will be more easy and cheaper so everything will evolve slowly and will get better AI will become at some point and I'm pretty certain that this will happen but I just don't know when it's so powerful that it offers a real-time translation and like in Star Trek for example with that Universal translator we use a device called Universal translator it's like an alien dictionary with hundreds of languages programmed into it and it can learn new languages very quickly but it doesn't always work so the necessity to learn a foreign language to communicate with a different person were more and more diminished but this is still in the future and we don't know when this will happen like to have like a really real-time translation which takes into account contexts and all the other stuff and which is comfortable to use and which is also cheap there's still many obstacles which need to be overcome universally am I the only one who bother to learn a foreign language so then the question is it still a good idea to learn foreign languages first of all it's a good exercise to train your brain to stay yeah fresh and young also age doesn't prevent you from learning new languages just look at Steve Kaufman for example then it will give you a better perspective on the world because you'll be able to understand different cultures and different points of view you will become more tolerant that is really good to connect to people on a deeper level because when they notice that you put in the effort to learn their language they will perceive you differently and also speaking in your mother tongue yeah it's a different kind of emotion for the people so it definitely does something then it's actually fun and can be a nice Pastime so yeah it's just a great feeling to consume content in a foreign language and to exchange yourself with other people and to talk and chat and it can be awesome sometimes it can be also great but in general I would say yeah it's a really pleasant experience it will build your confidence that you can actually do stuff and then you can learn stuff because sometimes especially when we are not probably not that good at school like I've been for example it can be hard to find trust in yourself that you can actually yeah Master something or learn something language is a great tool to do that so it's a great activity and when you notice that by spending enough time with a language and by doing your exercises and stuff you can learn something you can Master something then we also recognize that this applies to other areas in life and it will have definitely positive effects on yeah your whole life then languages are an important part of our culture and they will also remain being that then it has never been so easy to learn a foreign language nowadays with all the tools available and the internet and all the information at your fingertips we can also talk to tutors from all around the world you can even learn a language that doesn't get spoken in your area if you want to if you are interested in that different culture and then the day that AI becomes so powerful that it could really replace um yeah speaking for a language might be still very far away we don't know that so yes if you ask me learning for a language is a great thing and you should definitely do it it will enrich your life I'm certain of it and then last but not least if you found the video useful please leave me a like and a subscription to the channel I do not miss anything going forward click on the notification Bell thank you very much for watching and see you the next time

Channel: Dustin Schermaul

Views: 1,414

Keywords: AI and language learning, will language learning become obsolete?, machine translation, rule based machine translation, statistical machine translation, neural machine translation, google neural machine translation, adaptive machine translation, timekettle, bigbong, skype-translator, deepl, google translator, trint, no language left behind, google pixel buds, earbuds for translations, future of language learning, timekettle wt2, does it still make sense to learn a foreign language

Id: V9SzJ8AjXe0

Length: 11min 21sec (681 seconds)

Published: Sun Oct 16 2022