How Close are We To Getting Machine Translation Perfected?

machine translation

Panic among translators is evident owing to the continuous development of translation technology. Some years back, the question was whether it was possible for machine translation to replace human translators. Today, we no longer have to talk about the possibilities. There is proof beyond a reasonable doubt that human translators will be phased out at some point. The big question for many is “when?” Is perfect machine translation drawing closer or do we still have time to wait? Well, machine translation may be perfected sooner than you think. In fact, in some quarters, machine translation is already taking root with just a few hiccups. Let’s first explore the leading translation technology to be able to understand the strides made in translation.

Understanding Neural Translation Machine Technology- NMT

Over the years, there has been the development of many machine Translation technologies. However, most of the technologies had many shortcomings and were deemed unreliable and could not be trusted to take control of the entire process without human intervention. With the Neural Translation Machine Technology (NMT), things are much different. NMT is a technology that is based on learning algorithms, which are constantly receiving data and developing along with human evolution. Wikipedia cites that NMT can predict the likelihood of a sentence by modeling all the words in a sequence. In the past few years, most language translation service providers (LSPs) have largely benefited from NMT technology. This technology has improved the speed of translation, which in turn lowered the cost of translation. However, the elephant in the room is whether this technology can replace human translators and how soon?

Effectiveness of NMT In Perfecting Machine Translations

If you are a translator who is worried about the future of your job, NMT should give you a reason to start looking for an alternative job. Although we are not yet there, the prospects with NMT are very promising. The good news to traditional translators is that the translations that NMT produces still need human intervention. One Hour Translation, one of the leading translation service providers, is one of the companies that fully use NMT. According to the One Hour Translation team, they have to rate the performance of each NMT engine and add to the database on a quarterly basis. Do not forget that the NMT engines are developed on a logarithm that is continuously learning through data. Although the engines may not be perfect today, they are developing at a first rate and will be able to produce translations that are near perfect in the near future.

What Are The Shortcomings Of NMT

Since NMT is the leading technology in translation, it is probably the to effect machine translation. To understand the shortcomings of NMT, we must look at translation as an entire process. Translation involves LSPs and individual business that want to attain translation. This means that technology an NMT server must facilitate communication between the client and the translation service provider. If this process is to be effective, the NMT system should be able to handle multiple translation processes simultaneously. Unfortunately, while NMT technology is able to handle certain aspects of translation effectively, running multiple translations simultaneously does not give the system room to perfect in one area.

To handle such shortcomings, several NMT engines are used to offer different types of translations. It has been realized through human assessment that, some engines are effective in offering one translation than others. For instance, a certain engine may offer near perfect English to French translation but may not display such effectiveness when translating French to German. For this reason, it may take some time before the human translators are able to ascertain which engines function properly at what translation. Once this is established, we will be ready to have the NMT technology take over translation. Even if the technology may have some shortcomings, the number of human translators needed to assist the machines may be very limited.

The Effect Of Machine Translation So Far

According to One Hour, Translation’s CEO, over 40% of all translations in the world will be done using NMT by 2022. Although this is a personal opinion, it counts considering that over 30% of all translations already rely on NMT. Although NMT translations still rely on human support, the case may change sooner than you can imagine. While talking to Forbes Magazine, Ofer Shoshan explained that the level of dependence on human-assisted translation has reduced over the past 3 years. As at today, NMT has already impacted over 600, 000 linguistics through over 21000 translation service provision agencies.

What Next For Translators

If this article speaks to you as a translator, you might want to think that it is all over. Thankfully, we are not yet at a stage where machines can do everything independently. Even though we have technology that can offer direction in meters, tourists still need tour guides. In a nutshell, although we may soon have technology that can offer perfect or near perfect translations, it does not mean human translators become obsolete. There is an obsession with technology taking away jobs that belong to humans. The reality is that it does not have to be either technology or human; it can also be technology and humans. Humans should be ready to work alongside technology for a better planet. If technology can help humans get better translations faster, there is nothing wrong with giving it chance After all, faster translations simply mean quick execution of office process and increased overall productivity.

Conclusion

To answer the question of how close we are to perfected machine translation, we have to look at the development journey of translation. As things stand today, NMT technology seems to have an upper hand. However, the difference between NMT and other translation technologies is that it evolves with human language. With NMT technology, perfected machine translation may happen as soon as two years from now. However, this does not mean, human input will be obsolete in terms of translation. For business purposes, human input is still an integral part of translation and will always be.


Add Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

How Stacy Keibler Achieved a Net worth of $4.5 Million
SEM
How Search Engine Marketing (SEM) & TV Media Work In Tandem
Dean Foods CEO Ralph Scozzafava
10 Things You Didn’t Know about Dean Foods CEO Ralph Scozzafava
Laurance S. Rockefeller
10 Things You Didn’t Know about Laurance S. Rockefeller
Discover It card for Students
The 10 Best Credit Cards for People with No Credit
Texas Instruments
Why Texas Instruments is a Solid Long-Term Dividend Stock
Is Getting a Credit Card in College a Smart Move?
Crypto Investing: It’s Time for Smart-er Contracts that Encourage Innovation
Mixed Reality Technology
What is Mixed Reality and Where Are We With It?
5 Myths About Custom Mobile Applications
Seamless Virtual AI Assistant
How Close Are We to Seamless Talking AI Assistants?
Wearable ECG
How Will Wearable ECGs Affect Our Future?
10 Reasons to Stay at Le Meridien New York, Central Park
The Prime Rib
Why The Prime Rib is One of Philadelphia’s Finest Steakhouses
Barclay and Prime Philadelphia Steakhoues
Why Barclay & Prime is One of Philadelphia’s Finest Steakhouses
Juniper Serra Museum
10 Reasons to Visit the Junipero Serra Museum in San Diego
2004 Ferrari F430 Scuderia
A Closer Look at The 2004 Ferrari F430 Scuderia
1991 Ferrari TestaRossa 512 TR
A Closer Look at the 1991 Ferrari Testarossa 512 TR
The 1987 Ferrari F40
A Closer Look at The 1987 Ferrari F40
2000 Ferrari Rossa by Pininfarina
A Closer Look at The 2000 Ferrari Rossa by Pininfarina
Breitling GMT Light Body
A Closer Look at The Breitling Bentley GMT Light Body
What to Watch For: A Collector’s Interview
A Closer Look at the Breitling Bentley Flying B No. 3
2019 Breitling
Benefits of Authorized: Avoiding the Grey Market