Natural language processing machine translation

These companies either grew organically or through mergers and acquisitions. The applications of machine translation are also wide; in such fields as military and defense, healthcare, finance as well as technology. Most companies that make use of machine translation apps are motivated by the potential of growth in the market size and share as well as the profits and revenues involved. By 2024, market revenue is expected to be at 24 percent and reach $1. 5 billion. 1 Business Applications of Machine Translation 7 2. 2 Consumer Application of Machine Translation 8 SPECIALIZED MACHINE TRANSLATION COMPANIES AND SERVICES 8 3. 1 SDL Government 8 3. 2 Systran 9 3. 3 Lingua Custodia in Finance 11 3. 2 Failure to Comprehend Creative Language or Context 21 4. 3 Inability to Solve Ambiguity 22 CONCLUSION 22 REFERENCES 24 NATURAL LANGUAGE PROCESSING: MACHINE TRANSLATION 1. 1 Introduction Natural language processing has continually gained attention of representing and analyzing the human language computationally.

It makes use of computational techniques with the intention of learning, comprehending as well as producing the content of human language. According to Hirschberg and Manning (2015), the previous computational methods to language research concentrated on systematizing the scrutiny of the linguistic construction of language as well as developing basis technologies like speech recognition, machine translation and speech synthesis among others. With increased computational activities becoming more mainstream as more opportunities are availed by the internet to the global and multilingual community, R&D or research and development in Machine Translation is continually growing at a fast rate. There are varied kinds of Machine Translation obtainable in the market today and the most commonly used include; Rule-based Machine Translation (RBMT), Statistical Machine Translation (SMT) and Hybrid Systems that combine both SMT and RBMT.

1 Application of Machine Translation Machine Translation has evolved considerably over time, particularly in relation to the level of accuracy in output. The figure below shows the performance of Machine translation in relation to accuracy of translation of such languages as French, Chinese and Spanish to English or vice versa. Source: Madhavan, R. Even though such solutions are mostly automated, they are still highly dependent on human translators for both the post and pre-editing processes. Some of the fields that authorize machine translation solutions that are domain-specific include: Finance, Healthcare, E-discovery, Government, Military and Defense, Software and Technology, E-commerce and Legal. 2 Consumer Application of Machine Translation Such applications of machine learning execute instantaneous conversion for image, audio and textual files from an original language into a different language.

These are commonly generic, that is, not specific to a particular domain (non-domain specific) although with high accuracy of translation. Such applications are commonly light-weight, wearable devices or cloud-based apps that are commonly trained on crowd-sourced data. 3 million in 2017 and in 2018; it made a 13 percent increase in revenue, estimated at $417 million (Faes, 2019). There are about 20 players in this industry and the main competitors for SDL are Adobe, SYSTRAN, Sitecore, Wordbee, LionBridge and XTM International even though SDL ranks top in this field (Owler, 2018). The company acquired Donnelley Language Solutions in 2018, a company that offers translation, voice recognition, transcription and copywriting services. It also acquired Alterian, Language Weaver, Calamares and Bemoko in 2012, 2010, 2011 and 2013 respectively and has since recorded major profits. It is operational in about 38 countries and has 1500 customers (SDL, 2018).

9 percent, 8. 9 and 8. 9 percent of the revenue even though revenues declined during the early 2000s. The annual revenue has been estimated at $27. 6 million annually and the 2018 revenue is as shown below. Systran makes use of highly experienced and trained linguists and engineers to compete effectively and eventually surpass power players such as Google in the race to design a world that is truly linked without division based on language. 3 Lingua Custodia in Finance Lingua Custodia is the only machine translation company that specializes in translation of financial documents like fund prospectus, fund factsheets, fundamental information documents for investors, fund yearly reports and portfolio management annotations among others. From the company’s website, Lingua Custodia was launched in France, Luxembourg in 2011 and has been collecting financial linguistic information for over 7 years now and offers various levels of services to clients (Lingua Custodia, 2019).

VERTO is the major translation tool for the company and this tool is fully customizable with a particular ability to learn from the texts that had been previously translated by users. The company’s website fails to clearly state its customers but it states that it meets the needs of any professionals who need their financial documents translated. com/c/lingua-custodia/429669448 The reasons for this decline are unknown. In 2018, it was ranked the second best in the Fintech awards due to its innovative solutions in machine translation in the financial sector (Lingua Custodia, 2018). 4 Canopy Innovations – Healthcare Canopy Innovations is a digital health firm that was established in 2010 and headquartered in New York City. It has been operational for 19 years and its main motivation is to bridge the cultural and linguistic barrier between caregivers and their limited English proficiency or patients that are non-English speaking (Canopy Innovations, Inc, 2019).

Canopy Speak is an app for medical translations that runs on a phrase archive. 5 Consumer Machine Translation Applications 3. 1 Google Translate In the consumer market, Google’s Google Translate is the principal player. The capabilities of its real-time translation currently include speech, image (of words) and text, all packaged into one platform of a cloud and mobile app service. Google Translate was launched by Google in April 28, 2006 and has been in operational for 13 years. According to Greene (2016), Google Translate is made possible by the enormous trove of data that Google has as well as the statistical techniques match n-grams from a particular language with plausible n-grams to another. This means that there are several players in the field, about 100 in total. Google Translate also grew organically and in 2014, it acquired Word Lens, an app that can translate language using the phone’s camera.

Its financial statement is not available but with over 200 million users, it can be estimated to more than one million dollars and is mainly driven by revenue from the market. 6 Real-Time Text-to-Text Translation using NMT 3. 1 Facebook Translate Facebook has made various experiments on machine translation over the years. (May 17, 2019). Machine Translation – 14 Current Applications and Services. Emerj. Retrieved from https://emerj. com/ai-sector-overviews/machine-translation-14-current-applications-and-services/) Facebook acquired Jibbigo in 2013, a speech translation app. Skype began using the Microsoft Translate technology in 2017, a technology that operates on neural networks in order to precisely translate audio conversations into 18 languages as well as textual conversation in over 60 languages in a better manner. The Real-Time Speech-to-Speech translation field has more than ten players, among them being Mate, Google Cloud Text-to-Speech and EartthChat.

Skype Translate was obtained through acquisition by Microsoft for $8. 5 billion in 2011, and is considered as a big firm. Microsoft has a total revenue of $125. However, Skype robotically records the voice calls (once it issues a suitable caution to the handler) while using the translate feature in order to enhance the precision of the speech-to-speech translation as well as observe the unique tinges of languages. 2 The ILI Technologies Corp The ili refers to a small, wearable and handheld voice translator that is built for the purpose of translating common and simple phrases for travellers. Even though it is marketed as wearable, this technology that is based in Tokyo and made by Logbar Company only applies a simple chain for a traveller to put on around the neck with the device looking like a pendant (Javelosa, 2016).

It operates offline and does not necessarily need the Internet to instantaneously translate speech. It is presently only accessible in four languages namely Mandarin, Spanish, Japanese and English. 2 million in the first, second and third quarters of 2018 as shown below. ILI was acquired by SAI Global in 2006 and is said to have boosted annual earnings of the company to more than $70 million (SAI Global, 2006). ILI is considered a big company that is driven by revenue and profits and has been operational for 17 years. MOTIVATIONS FOR USING MACHINE TRANSLATION From the apps discussed above, it can be noted that there are various firms that make use of machine translation, with Microsoft, IBM and Google being the most recognized protagonists because of the role that they play in the consumer market.

Other companies that commonly use the app include Moravia, Systran and LionBridge among others. Why are the big tech companies so obsessed with machine translation? Prodigious Company. Retrieved from https://www. translateplus. com/blog/big-tech-companies-obsessed-machine-translation/ Military and defence are said to take up the largest share in the market for machine translation followed by electronics and IT. Madhavan (2019) claims that by 2024, the estimated market size of the industry for NLP is expected to be $2. Additionally, in an incident where confidential data is involved, it may be difficult to know just how safe the use of a third-party machine translation technology could be. 2 Failure to Comprehend Creative Language or Context Machine translators are incapable of comprehending creative language use of context. For instance, slogans, metaphors and puns in a word may make little sense when converted into another language (Mimnagh, 2018).

On the other hand, a human translator would be able to recognize certain aspects in which literal translation is not possible and even find the most appropriate alternative. 3 Inability to Solve Ambiguity Machine translators are normally unable to solve problem or ambiguity whereas a human translator would. The machine translation field is expected to grow significantly, at about 24 percent, hitting $1. 5 billion by 2024. Furthermore, the potential growth in market size for machine translation apps encourages new players and increased advancement in the area. Most of these players such as Google and IBM apps have grown organically while others such as Microsoft’s Skype have developed through mergers and acquisitions. Even though advancement has been made in machine translation, there are still various weaknesses that human translation can eliminate.

Websites: Berger, Y. (October 22, 2017). Israel Arrests Palestinian Because Facebook Translated 'Good Morning' to 'Attack Them. ' Haaretz. Retrieved from https://www. com/company/canopy-health Canopy Innovations, Inc. Canopy Innovations. Retrieved from https://angel. co/company/canopy-nyc Greene, L. (February 2, 2016). net/think-tank/articles/everything-you-ever-wanted-to-know-about-google-translate-and-finally-got-the-chance-to-ask Hern, A. (October 24, 2017). Facebook translates 'good morning' into 'attack them', leading to arrest. The Guardian. Retrieved from https://www. Meet ili, the First Real-Time Wearable Translator In The World. Forbes. Retrieved from https://futurism. com/meet-ili-first-real-time-wearable-translator-world Kravariti, A. (May 24, 2017). zoominfo. com/c/lingua-custodia/429669448 Lingua Custodia. Lingua Custodia. Retrieved from https://www. lhoft. Pros and Cons of using Machine Translations. International Translations Limited. Retrieved from http://www. itltranslations. com/technology/the-pros-and-cons-of-using-machine-translations/ Mitchell, J. ibm. com/blogs/watson/2018/07/improving-the-accuracy-speed-of-translations-with-neural-machine-translation/ Owler.