Browse Prior Art Database

Method and system for learning based automatic input method language recommender

IP.com Disclosure Number: IPCOM000226314D
Publication Date: 2013-Mar-27
Document File: 5 page(s) / 107K

Publishing Venue

The IP.com Prior Art Database

Abstract

This disclosure discloses a method and system enables the automatic input language switching dispensing user's manual effort. Such a language switching engine works in the cases like using different languages when communicating with people from different nations, when using different softwares and applications, and when the editing document is not written by the current input language, just name a few. Specifically, this engine collects the data from users' daily communication and document editing, and it is trained by these data to enable its predictive capability to automatically switching between different languages as needed.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 5

Method and system for learning based automatic input method language recommender

The disclosed method and system include two stages: offline and online. In the offline stage, the daily context information generated by users are feeded to a data-driven predictive model, and this model is trained by these data to learn the input language preference for a specific user in different contexts. In the online stage, the trained model is performed to predict the input langauge preference given current context information, such as who the user is communicating with, what software

application is used, and what is the time (day or night, entertainment time or office time), etc and automatically switch to the preferred language without requiring user's manual effort.

As an embodiment, in the offline model learning stage, two factors, the communication/editing platform and the person who are communicated with can be considered and modeled appropriately.

In particular, the communication platform (typically a software) can be embodied by the current instant messaging tools which help sending the messages between two or more users. And for the editing task, the document profile including document title, text language, document topic etc can be knwon/extracted. And the translator can be regarded as a good example when the user

reads one language, he/she needs to put the language wording into the translator and let the translator translate it to another language. So automatically switching to the currently reading language is welcomed. Also, for the connected communication case, the specific applications is a good indicator, for instance, if the website is natively produced in China, Chinese is automatically

welcomed.

And another factor is the person who is/is going to be communicated with. In this case, the static profiles associated with the two/multi-party use...