Hi User,
For the past 6 months, I have been actively developing v2.0 of AA Reader. In this new version, the signature "context-aware word sense engine" will be a lot more powerful. In the previous versions, if you select Chinese as your mother tongue, when you tap on a word, the system will try to figure out the type of the word (noun, adj, verb etc...), then shows the respective Chinese translation of the word. This indeed improves the reading experience by a large margin, but it contains some drawbacks:
1. Many English words have very distinct senses, e.g. bank, clear, run, book, table etc...
2. Only works on English-Chinese translation
With the help of the latest state-of-the-art NLP (natural language process) technology, in the upcoming new version of AA Reader, I will try to address these drawbacks by completely replacing the existing translation engine.
I have put up a test web page to illustrate this feature:
In this web page, you can select a word and select a sample sentence, or key in your own sentence then press the submit button, the system will tokenize the sentence. Then you can select a word to explain (currently only a dozen words are supported). The system will present the most likely sense of the word.
Example: The man who robbed the bank was later seen at the river bank.
The first bank will give you "financial institution....", the second bank will give you "sloping land...".
With this new technology, when you see a new word in a book or article, all you will be presented is the contextual sense of the word in that particular sentence instead of the full dictionary page which can be very lengthy. This will give you a better reading experience while also improving the efficiency of learning new words.
This test server is a baseline machine located in Singapore, it may be a bit slow to access from your place.
As the development is only at the infant stage, I need your help to test out the system, give me feedback, so I can make the product even better, thank you!
Regards,
AA Reader
网友评论