Imagine if you have an online shopping website and you have a variety of customers. Not only from Indonesia, but you also have customers abroad. It looks really good, right? But do your customers from outside the market understand your language? If we refer to online shopping websites such as eBay and the like, they not only have customers from the United States, but they, of course, have customers outside their home country, whether from Japan, Singapore, or Indonesia.
In this case, Google mentioned the surprising fact that 40% of global customers do not recommend buying online goods from websites that do not share their language with 51% of internet users not using English or Indonesian as their main language. Of course, you will miss a big opportunity and also lose a market that has the potential to generate profits for your business.
In this case, the company is faced with a choice that is not easy. Here are some of the most frequently encountered cases:
Takeaways:
Quality translators have limitations and are expensive
Machine translation that has standard quality. Fast but not accurate
The risk of translating through oneself is risky and inconsistent
If we look again, to answer this translation challenge, you certainly need all three, and traditional translation methods must always meet standards, so it is difficult to adapt. This problem is, of course, not only translating word problems so that customers understand what you mean but also connecting between customers so that your customers understand the context you explain.
To provide a more in-depth explanation, we will discuss one on one sessions with Cloud Engineer, Aldi Fianda Putra, who will explain how Cloud AI Translation technology is able to translate languages with high accuracy and the workings behind the technology.
What are the latest updates to Google Cloud Translation AI?
There are many things updated in Translation AI on GCP, including:
First, the thing that has been improved is language coverage, which includes around 189 languages, and increased translation accuracy
Second, the translator is more adaptive and intelligent, for example, we can adjust the tone and style of the translation between 5 to 30000 examples for more precise results
Third, Translation AI itself consists of 2 types, namely basic and advanced, the advanced type offers various translation model options according to our translation use case, for example, we can use the NMT model for common languages or use the adaptive model for real-time customization
Finally, the Translation AI service also provides evaluations to help assess the quality of the model based on our specific criteria or use case
What are the benefits of using the latest Google Cloud Translation AI update?
With the Translation AI update, many benefits can be gained, one of which is the expansion of language coverage, which now includes 189 languages, including less commonly used languages such as Cantonese and Fijian. This allows businesses to reach a wider global audience with fast and accurate translation quality. Additionally, the adaptive feature allows us to train the model according to our specific context, such as industry terms or a certain tone, and all this can be done without needing to write any program code at all.
How can the latest Google Cloud Translation AI update improve translation quality?
Translation AI improves translation quality by using Neural Machine Translation (NMT) models that produce more accurate and natural translations. Adaptive features enable translation adjustments based on context, such as tone or technical terms. Users can also upload custom glossaries for terminology consistency, while AI-based evaluations such as MetricX and COMET help continuously monitor and improve translation quality.
What are some use cases for the latest Google Cloud Translation AI update?
Translation AI has a variety of relevant use cases across various industries. Perhaps the most relevant example is multilingual customer service, Translation AI can be used to translate messages in real-time in contact centers or chatbots, enabling interactions without language barriers. In the e-commerce sector, product descriptions and customer reviews can be translated to reach global markets. Apart from industrial examples, for simple cases we can also use this AI translation to translate the content on our website pages.
How is Google Cloud integrating user feedback into its latest Translation AI feature?
By using Translation AI, we can integrate input through several mechanisms. For example, we can customize the tone and style of the translation by providing translation examples ranging from 5 to 30,000 examples. Apart from that, we can also upload specific datasets according to our needs, such as datasets containing medical, scientific, technical and other terms, which will be used to train adaptive models so that translation results are more accurate and contextual. All of this can be evaluated to measure how well the model meets our specific needs, and this process can be done through the Vertex AI service, which allows rapid prototyping without the need to write program code
Google tracks location data even when users turn service off(Opens in a new browser tab)
Through Google Cloud Translation AI technology, businesses can now reach global markets more easily, deliver better communications, and provide an optimal customer experience. It’s not just a matter of translating languages but also establishing more optimal relationships with customers.
Source: Sharing the latest updates to Google Cloud’s Translation AI | Google Cloud Blog
Discussion about this post