Translate PDF Files and Office Documents

This seems to be a recent change: Google Translate can now be used to translate PDF files and Microsoft Office documents. Google first converts the document to the HTML format and then it displays the translated HTML file.

The nice thing is that Google converts the documents on the fly even if they haven't been indexed by Google. Just paste the address of a document in Google Translate's textarea and wait until the translation shows up. Unfortunately, the service is not well-suited for long documents: it only translated the first 9 pages of a PDF file.

7 New Languages in Google Translate

Expect 7 new languages in Google Translate soon: Albanian (9-13 million speakers*), Estonian (1.1 million speakers), Galician (3-4 million speakers), Hungarian (15 million speakers), Maltese (around 400,000 speakers), Thai (60-65 million speakers) and Turkish (63 million speakers). Many of the comments from a previous post requested Turkish, so Google listened to the feedback.


Another change is that the English dictionary has been improved significantly and it includes synonyms, antonyms, pronunciations, detailed definitions and examples from Collins COBUILD Advanced Learner's English Dictionary.


Update (a day later): As previously announced, the new languages have been added to Google Translate.


* Source: Wikipedia.

Translate Google Maps Reviews

Google Maps has a new option that automatically translates reviews written in a foreign language. This is especially useful if you are planning to visit a different country and most of the reviews are not available in a language that you speak. Here's a list of results for [bistro, paris, france] in the English interface of Google Maps and a search for [pizza ny] in the French Google Maps.


Similar options are available for Picasa Web comments, YouTube captions and many other services, thanks to the easy-to-use API for Google Translate.

It's worth mentioning that Google Translate's homepage has been redesigned and you can now translate text and web pages from the same box. Google Translate includes a new option that swaps the selected languages and the translations for short texts have permalinks.

Picasa Web Albums Translates Comments

Believe it or not, Picasa Web Albums is a big success outside US and, since the interface is available in 38 languages, many people post comments in their native languages. Click on a featured photo and you'll certainly see at least a comment written in Spanish, Italian, Chinese, French or other language.

Google's photo-sharing service started to use Google Translate API to automatically translate comments in your interface's language. The translation is displayed below the original comment.


Of course, the translation is not perfect and Google didn't do a good job at properly disclosing that it's an automatic translation (there's a "Powered by Google Translate" tooltip when you hover over the translation). This is yet another integration of Google Translate in Google's services, after translating Google Reader subscriptions and YouTube's search results. What other services would you like to integrate machine translation?

Translate Feeds in Google Reader

Just because you don't speak a certain language doesn't mean you shouldn't read interesting articles written in that language. After increasing the number of supported language pairs and providing an API, Google Translate is rapidly integrated in many Google services. YouTube started to translate search results and closed captions, Google Talk provides translation bots, Google lets you search content in other languages and now you can subscribe to feeds in foreign languages using Google Reader.

After subscribing to a feed in Google Reader, click on "Feed settings" and select "Translate into my language". Google detects the feed's language and it translates the content into Google Reader's interface language (you can be changed it from the settings page).


Some Google-related sites in other languages than English:
Zorgloob (French)
Google Watch Blog (German)
OJObuscador (Spanish)
Google Discovery (Portuguese)
underGoogle (Portuguese)

When Google's machine translations improves, the next step could be translating content into your language without having to select an option. In the meantime, you can expect to find Google Translate integration in Gmail, Google Docs and other Google services.

{ via Google Reader Blog }

Machine Translation and Speech Recognition at Google

Google's technologies for automatic translation and speech recognition already have visible results: you can translate texts in 35 languages at translate.google.com or use voice to find a local business with GOOG-411, but Google intends to expand their use. You should be able to translate an email written in a foreign language or find answers to simple questions by voice.

Mike Cohen, who leads Google's speech technology efforts, and Franz Och, machine translation researcher, chat with Alfred Spector, VP of Research and Special Initiatives at Google, about two technologies that might seem unrelated to Google's core competency. Both statistical machine translation and speech recognition are search problems and Google's computer infrastructure can process large amounts of data that are needed to build language models. Another big advantage for Google is that it has popular services that generate a lot of useful data.

"When we first created GOOG-411, we had no speech data. Because we had so much query data here at Google (textual queries that people had typed to Google Maps), we could already train a pretty good language model. Now, obviously, text is a little different than speech and now that we've also trained on speech, we have better performance than we had back then, but even out of the box we could get good performance on that problem because we had so much textual data," says Mike Cohen.

More Languages in Google Translate

Google Translate added 11 new languages: Catalan, Filipino, Hebrew, Indonesian, Latvian, Lithuanian, Serbian, Slovak, Slovenian, Ukrainian and Vietnamese. Google's machine translation service now supports 35 languages and you can use it to translate text between any combination of languages.

In most cases, Google uses English as an intermediary language, so when you translate a text from Indonesian to Vietnamese, Google translates the text to English and then it translates the result to Vietnamese. You'll get the best results when one of the languages is English, since Google needs a single translation.


"Most state-of-the-art, commercial machine-translation systems in use today have been developed using a rule-based approach, and require a lot of work to define vocabularies and grammars. Our system takes a different approach: we feed the computer billions of words of text, both monolingual text in the target language, and aligned text consisting of examples of human translations between the languages. We then apply statistical learning techniques to build a translation model," explains Google.

One of the advantages of this approach is scalability: if Google finds enough parallel text to create a good translation model for a language, it will be added to Google Translate. When Google licensed Systran's technology, Google Translate was only able to translate between English and French, German, Spanish, Italian and Portuguese, but this has changed when Google developed its own translation technology. Microsoft followed suit and Windows Live Translator switched from Systran to Microsoft's machine translation system.

As Microsoft notes, it's important to keep in mind that "automatic translation enables you to understand the gist of foreign language text, but is no substitute for a professional human translator if fluency is required," at least not yet.

Google Translate OneBox

Google has a new OneBox for quick translations. You can just search for "translate", followed by a word or an expression and the optional "into English". For example, you can search for [translate désormais] and Google shows the entry from an automatically-generated bilingual dictionary.



Unfortunately, you can't use the search box as a shortcut for Google Translate because full texts aren't yet translated. The bilingual dictionary is only available for the following language pairs: English <-> French/Italian/Spanish/Portuguese/German/Russian/Chinese/Korean/Hindi. It's interesting that Google shows results from Google Image Search next to the translations.


{ Merci, TomHTML. }

Translate YouTube's Search Results

YouTube tests a new option that translates the search results into your language using Google Translate. Now that you can independently select the interface language, it would be nice if the user-generated content was available in the same language.

YouTube didn't translate the titles and descriptions for all the videos, but it decided to use Google Translate's JavaScript API to perform the translations when they are needed, much like Twitter Search. If you click on "translate results into my language", YouTube sets a preference and all the search results will have titles and descriptions in your language, assuming it's supported by Google Translate.

The translation API offered by Google can be easily integrated in web applications, especially that it can now preserve HTML markup. I expect to see a lot of applications that take advantage of the API, from webmail services like Gmail to browsers and IM clients.



In other machine translation news, Google Dictionary has a new interface and it now supports more language pairs. There's also a section for definitions found on the web.


{ Thanks, Aleksandr. }

Google Translate Becomes the Best Free Online Translator

Google Translate's coverage has been expanded dramatically. It now supports the translation between any of the following languages: English, Arabic, Bulgarian, Chinese, Croatian, Czech, Danish, Dutch, Finnish, French, German, Greek, Hindi, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Romanian, Russian, Spanish, Swedish (the new languages are shown in bold). From 26 language pairs, Google Translate now supports 506 language pairs and becomes the most comprehensive online translation tool available for free.

Obviously, the translation is far from being perfect or even coherent, but it's a great way to understand the central ideas from a text. Now that Google Translate supports so many languages, it's not hard to imagine that you'll be able to read almost any web page in your language and maybe any application will be able to use Google Translate's APIs to speak your language.

"Most state-of-the-art, commercial machine-translation systems in use today have been developed using a rule-based approach, and require a lot of work to define vocabularies and grammars. Our system takes a different approach: we feed the computer billions of words of text, both monolingual text in the target language, and aligned text consisting of examples of human translations between the languages. We then apply statistical learning techniques to build a translation model. (...) Automatic translation is very difficult, as the meaning of words depends on the context in which they're used. While we are working on the problem, it may be some time before anyone can offer a quick and seamless translation experience," explains Google Translate's FAQ.


{ via Page2RSS }