Neural Translation Machine Has Made Literature Translation By A Machine Possible

A recent research done by Neural Translation Machine suggests that the technology is now capable of translating Literature with an efficiency of 25%. Though this number may not seem like a lot, the complicated nature of translating literature is an insurmountable task indeed. Thus, it is a remarkable achievement marking the progress of our technology.

Platforms like Google translate and Skype translate are the most famous users of Neural Translation Machine. Though the fact about the real technology which is used for the working of NMT remains a puzzle for the users.

As more development continues to take place in machine translations, eventually it will become easier for the users to understand the working of the NMT. The 25% flawless rate suggests us that technology is moving in the right direction with regards to machine translation.

Neural Translation Machine

Neural Translation Machine Hits 25% Flawless Rate

Natural Spoken word is a very natural element. The toughest test for the machine translation is to understand the true meaning of the words used by us. Capturing the same meaning in a different language is far-fetched.

As an example, the word ‘set’ can be a verb, a noun or an adjective with various meanings and it depends on how and where you are using it in a sentence. Similarly, there are more such words which are used in different manners and ways. Therefore, it is important for the NMT to understand the realize the meaning of the words depending on the sentence.

The aim of the Neural Machine Translation is to figure out the meaning of the words and phrases by assuming the order of the words and applying the texts through deep learning.

This will improve the decision making of the technology and then it can decide when to use the word ‘set’ as a noun, a verb or an adjective. Once the decision making gets better, the accuracy of translation is bound to improve even more.

The Use of NMT in Novels

Using the Neural Translation Machine or NMT to translate parts of Literature is a complicated matter. However, Dr. Antonio Toral from the University of Groningen and Professor Andy Way made this possible via their research project.

NMT in Novels

Following are the twelve well-known novels Neural Machine Translation uses for its translational activities:

  1. Auster’s Sunset Park – Released in 2010
  2. Collins Hunger Games #3 – Released in 2010
  3. Golding’s Lord of the Files – Released in 1954
  4. Hemingway’s – The Old Man and The Sea – Released in 1952
  5. Highsmith’s Ripley in the Water – Released in 1991
  6. Hosseini’s A Thousand Splendid Suns – Released in 2007
  7. Joyce’s Ulysses – Released in 1922
  8. Kerouac’s On The Road – Released in 1957
  9. Orwell’s 1984 – Released in 1949
  10. Rowling’s Harry Potter #7 – Released in 2007
  11. Salinger’s The Catcher in the Rye – Released in 1951
  12. Tolkien’s The Lord of the Rings #3 – Released in 1955

The duo of Dr. Antonio Toral and Prof. Andy Way trained the Neural Machine Translation to translate the novels from English into Catalan. Thereafter, they matched those results with the Phase-based statistical machine translation (PBSMT) system with the same work of the Literature.

The Rise of The Machine Translation

The reports of the research project disclosed by Dr. Antonio Toral and Prof. Andy Way indicates just how far Machine translation has come in recent years. This is a big achievement to the advancement of our technology and efficiency in the development of the machine learning used in this process.

A decade ago, it was impossible to imagine that the algorithm can be even used for translating the Literature and other inventive processes. Indeed, the 25% flawless rate is not good enough for the real world scheme. However, it does indicate the gradual rise of Machine translation.

If the progression can be made so that the machines translate the documents to about 50% accuracy for the human translators, then the technology is going to have a big impact on how we deal with the language barriers.

Ultimately, the reason for putting the Neural Machine Translation to a test of this kind is to put it through some of the most challenging situations and check how well it progresses.

The results of the tests were superb. This is a reason to be optimistic that the machine translation will one day play an important factor in how the businesses and people communicate with each other.