Browse Prior Art Database

Interests view for unread messages based on cognitive computing and emotion detection

IP.com Disclosure Number: IPCOM000248900D
Publication Date: 2017-Jan-20
Document File: 7 page(s) / 90K

Publishing Venue

The IP.com Prior Art Database

Abstract

This invention discloses a method to pickup user's interested messages based on user response and emotion detection. The invention gets user's basic interested areas, according to user's profile and social network activity. For each social network message in the group, including text, link, picture, video, audio, interactive element(e.g. red envelop) and so on, it extracts keywords and get the category, topic information. When user is watching the message, system detects the user's interest level of such messages, based on detecting user's emotion change, eyeball focus, average reading speed and response. It trains the cognitive computing system with history message and relative interest level, when new message comes, it can predict the interest level of new message.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 38% of the total text.

1

Interests view for unread messages based on cognitive computing and emotion detection

People usually have different groups in social network. Although the group is created on certain purpose, the topics discussed in this group may be various, E.g. A group created by the teacher, for all parents of his students, the topics may include students homework, school activities, thank you from parents, or just casual talks. Another example is a group created by mothers to share raising baby related info, the topic may differ from group purchase to cute baby picture sharing.

User's interests change overtime, it may also differ in different user groups, even in same user group. E.g. the on sale messages broadcasted by friend A in the group is quite useful, so user often feel interested. But the on sale messages broadcasted by friend B is most likely to be just advertisement, so user usually does not spend time on it. Another example is about funny animation or interesting video. User just prefer some of them to others.

There may be more than one hundred messages in a short time, when there are many people in the group, or several people are talking with passion. If keeping the group new message notification on, it may be quite annoying , if keeping the notification off, user may not catch interested messages in time. And it is also very time consuming to check all unread messages one by one. This invention discloses a method to pickup user's interested messages based on user response and emotion detection. The invention gets user's basic interested areas, according to user's profile and social network activity. For each social network message in the group, including text, link, picture, video, audio, interactive element(e.g. red envelop) and so on, it extracts keywords and get the category, topic information. When user is watching the message, system detects the user's interest level of such messages, based on detecting user's emotion change, eyeball focus, average reading speed and response. It trains the cognitive computing system with history message and relative interest level, when new message comes, it can predict the interest level of new message.

The emotion and eyeball focus detection are resource consuming compared with other activities such as keyword and category extraction, so it is triggered by certain criteria: E.g. Battery over 40%, user is reading more than 50 unread messages.

When there are many unread messages, the invention creates a link for interests view, so that user can check interested messages quickly. And user can also select to auto-remove the uninterested messages or just mark them as read. The messages can be classified by category for quick checking.

Claim points: A method pickup user's interested messages based on user response and emotion detection. Interested level is calculated based on detecting user's emotion change, eyeball focus, average reading speed and response. Trigger emotion and eyeball focus detection in...