InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Leveraging Semantic Analysis for Notification Management

IP.com Disclosure Number: IPCOM000241338D
Publication Date: 2015-Apr-17
Document File: 4 page(s) / 84K

Publishing Venue

The IP.com Prior Art Database


Disclosed is a system that uses sentiment analysis via personal metric devices to manage electronic communications such that a user receives messages that are most appropriate for the current emotional state. For example, if a user is in a negative state, then the system determines that the user should receive positive messages; messages that might increase the negative state are put in a queue for when the user is in a neutral or positive emotional state.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 35% of the total text.

Page 01 of 4

Leveraging Semantic Analysis for Notification Management

As personal metrics devices continue to increase in popularity, communications systems can take advantage of increasing context around the user's physical state. In a simple example, a user's smartphone might disable notifications while the user is exercising. These types of capabilities are beginning to emerge in applications (apps) that integrate personal metrics with other devices. In the context of communication, systems also exist to try to integrate a user's state with activities. For example Google's* Mail Goggles require the user to solve a simple puzzle before sending an email late at night.

However, these solutions tend to focus on the initiator of the activity, and the Google Mail Goggles do not prevent a person from reading email while impaired. Additionally, a person in a bad mood may not want to be contacted by communication that can potentially put the recipient in a worse mood. In these cases, a method is needed to alter or delay communication paths to be considerate of the recipient's mood.

Current systems have rules that organize emails/chats/texts/messages according to time, recipient, or relationship hierarchy. For example, if a user is at work, the user can choose not to be notified of personal chats and texts from social networks or email. The user may be notified of emails/chats/texts from a spouse if there is an indication of emergency -- by either rule-based algorithms or semantic analysis. However, these systems do not automatically react to the user's state.

The proposed system manages notifications based largely on the receiving user's sentiment and combinations of the user sentiment with contextual hints. Mediums of notification can include text, email, smartphone, instant messaging (IM), voice messaging, etc. and take appropriate action based on that medium to react to the recipient's mood. As a core feature, user sentiment can be used to assess when personal notifications are presented.

Management of incoming communication is based on combined recipient and message sentiments. The system predicts the user's sentiment upon receiving the communication, and then takes steps to manage predicted sentiment. The system incorporates personalized learning of a user's sentiment reactions based on message content and sentiment state. This includes a combination of multiple sentiment factors such as the sender's sentiment and recipient's sentiment to manage communication. The system can also incorporate additional dimensions that can affect sentiment, such as requiring follow-up actions, long messages, and degree of emotion.

The main implementation steps follow:

1. Leverage a device that measures user's physical and derived mental state (bad mood, frustrated, exhausted, etc.)

2. Use sentiment analysis to determine the sentiment of communication being sent to the user, for example email positivity / negativity

3. Correlate the user's mood with the communic...