Reponses To Discussion

Reponses to discussion

[Name of the Writer]

[Name of the Institution]

(Question 1) Student 1 post: Salma Nawar

Machine translation is the translation of text by a computer, with no human involvement. It is used by many companies and organizations to adapt their products into various languages. Machine translation is there to make it easier to convert something from one language into another within minutes. 

There are various types of machine translation including: rule-based machine translation (RBMT), statistical machine translation (SMT), example-based machine translation (EBMT) and neural machine translation (NMT). RBMT works by skimming through a sentence to identify words and analyze their structure, then converting it into the language of your choosing. SMT works by training the translation engine with a large volume of bilingual texts, after which the system looks for statistical correlations between the source texts and the translations.  In EBMT a sentence is translated through an analogy. Last but not least, NMT uses neural networks that consist of nodes modeled after the human brain to fulfill it's translation duties.

Responses:

Machine translation is a program for translation of text to other languages by the computer. The student has explained the whole concept of Machine transition in an efficient manner. All the types of machines translation that are discussed by the student were defined in a precise way and in simple words which aid in the understanding of the concepts effectively. However, the weak point of his post is that it does not highlight the advantage of each type of machine translations

(Question 1) Student 2 post: George Torres

 Machine Translation, MT for short, is a systematic translation machine that outputs an automated translation from an input language. More precisely, it is computer software translator that takes one language such as English and translates it to another language such as French. There seems to be two main translation methodologies that are used in MT.

The first one is:

                             -Rule-based MT: Rule-based MT is a type of machine translation that is able to deliver high performance, high quality translations. Although this the ideal, the translations themselves are not ‘fluent’ and are very costly to perform. (systransoft.com). This leaves rule-based translation difficult to maneuver.

              The second one is:

                             -Statistical-based MT: On the other hand, statistical-based MT is basically the opposite of rule-based MT. It can produce fluent texts and the cost development is practical; however, the software does not know grammar nor delivers a ‘predictable’ translation (systransoft.com). In addition to rule-based MT, this software also comes with its downfalls.

              Overall both types of MT are good yet come with their imperfections. As suggested by systransoft.com, there is a need for a third type of MT that could deliver superior translations at low and practical costs.

Responses:

The student has given a brief introduction of Machine translation both in professional language by using words like automated translation and in simple language for general understanding. In the post, only two main translation methodologies are discussed in detail. I feel that other types of machine translations like neural translation are also important. And lack of mentioning of these type of machine translation makes this post weak and incomplete

(Question2) Student 1 post : Jackson Bjika

Neural machine translation is more important because it reduces post-editing effort by 25%, outputs more fluent translations, and “linguistically speaking it also seems in quite a few categories that it actually outperforms statistical machine translation

Everybody challenges to know how neural machine translation works because communication with all languages and all countries becomes easy than before it. Event in business pouches everybody on same international way to buy and sell.

Responses:

The advantages explained by the student are a bit vague. I am unable to understand the point student made about neural machine outperforming statistical machine translation. There are a few grammatical mistakes present like the student has used opening quotation marks only. The second point is also vague, and the meaning is not clear. Overall it is a weak post which lacks consistency and coherence.

(Question 2) Student 2 post : Salma Nawar

Neural machine translation (NMT) is a technology that has only become viable in recent years. It is one of the newest approaches to machine translation. It uses neural networks made of nodes that are modeled after the human brain. The nodes can hold single words or phrases and relate them to each other in a web of complex relationships. This is all based on bilingual texts that are used to train the system, similar to statistical machine translation. 

I agree with the statement "neural machine translation is the future." Neural translation is so advanced and it's a really great technology that we need to utilize. I can see it as becoming the main machine translation technique in the future because of how advanced it is. The fact that is similar to the human brain should tell you all you need to know

Responses:

It is a very strong post where all the aspects of the question are discussed in the right proportions. The student first explained how neural machine translation works. It makes it easier to understand the importance of Neural Machine translation. Based on this explanation the student than declared that he agrees on the notion that neural machine translation is the future because it is capable of beating the human brain. Another strong point of this post is that there is a steadiness in the flow of ideas and sentences.