Browse Prior Art Database

Method and apparatus for a low cost semi-automatic defense system against bullying in social networks.

IP.com Disclosure Number: IPCOM000242983D
Publication Date: 2015-Sep-04
Document File: 2 page(s) / 59K

Publishing Venue

The IP.com Prior Art Database

Abstract

Translation between languages of different geographical origins often provides a flawed contextual result, with the original meaning left intact but emphasis and nuance removed. This can be exploited in order to remove malice and harmful intent from a piece of text, in this instance, harmful messages on social networking sites

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 01 of 2

Method and apparatus for a low cost semi-automatic defense system against bullying in social networks.

Algorithms and decision engines for recognising intentionally offensive messages in social networks can be non-trivial and hence costly to implement. Part of the difficulty is the context of which the messages are uttered. Some intentions can become evident from keywords, as well from the structure of the message (assertive, passive, interrogative).

    Upon the principle that the least damage from an intentionally abusive message is when it is not understood by the target, an effective deterrent against an abuser could be provided by the situation when the offensive messages have been translated to an orthogonal language, then translated back to the original language it was uttered (for example, from English to Japanese, then back to English). In other words, the abusive message becomes lost in translation. The abuser, hence,

would get frustrated that his intentions are deflected. This principle could be implemented as a client side translator, which processes the utterances at low cost, updating conditional probabilities for patterns, keywords, contexts, and updating this knowledge to the server corpus repository, either asynchronously or in real time. Consider a generic description for a Social network site, comprising of:
a User: owner of an account in the Social network site. This could be the target of abusive posts/messages
an Audience: any number of users of the same network, that can read and post comments on the User space.
a thread/flow of messages by Audience members, regarding a topic(s) on the User space.

The preferred embodiment would consist of a client plugin in the User space,

which works as follows: The plugin parses each message M_Ai posted by Audienc...