As the number of participants in a group chat conversation increases, the problems of multiple topics and frequent topic switches become salient. These increase the burden on individual users to keep track of the conversation topics. These also lead to inefficiencies due to the disruption and interruptions caused by messages and notifications on topics that might be irrelevant or uninteresting to some of the chat participants. To address these problems, with user permission, this disclosure utilizes a machine learning model trained for automated detection of discussion topics in a chat. The detected topics are presented to the user to facilitate appropriate action based on current discussion topics or occurrences of topic switches. Model inference operations are performed entirely on the user’s device with specific user consent, thus preserving user privacy.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Feuz, Sandro and Deselaers, Thomas, "Improved Group Chat User Experience by Use of Topic Detection", Technical Disclosure Commons, (September 24, 2018)