Abstract
Emojis are a popular way to communicate emotions and gestures. As a type of visual art, the meaning of emojis is subject to individual perception and interpretation. The meaning and usage of emojis evolves over time. New emojis are introduced periodically. Discovering appropriate emojis can thus be difficult. This disclosure describes the use of machine learning techniques to automatically suggest emojis based on text input by a user. A machine learning model is trained over emoji annotations as well as usage patterns in a manner that adapts to changing meaning and usage of an emoji. Text entered by a user is mapped to emojis using the machine learning model. Matching emojis are ranked and provided as suggestions or autocomplete entries in the text composition interface. Emoji suggestions can be incorporated into a virtual keyboard, in a chat/messaging application, or in any other context in which emojis are used. The described techniques enable users to compose messages faster; discover and use new or trending emojis; select emojis that match the sentiment in their messages; etc.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Xin, Fuxiao, "Context Aware Emoji Suggestions Using Emoji Embeddings and Usage Pattern", Technical Disclosure Commons, (August 21, 2023)
https://www.tdcommons.org/dpubs_series/6163