2021:Submissions/Domain Specific Content Generation using Human Bot Collaboration
Speakers[edit | edit source]
Vasudeva Varma, IIIT Hyderabad, India
Praveen Garimella, IIIT Hyderabad, India
Abstract[edit | edit source]
The availability of Wikipedia content in regional languages, specifically in the Indian languages, is very low. For instance, Telugu has around 1% of the English Wikipedia. Several groups attempted the use of machine translation which resulted in poor quality content. Given this scenario, bot-generated articles using templates and structured databases proved to be efficient to engage the volunteers to enrich the regional content. In the pilot study, computer science students from the local region volunteered to use the framework to create close to 50K bot-generated articles in 7 different domains. Everyone else in the community enriched the bot-generated articles in the wiki way.
Topics: Education and GLAM, Innovation and Research, Languages, Our Stories
Session Outcomes[edit | edit source]
Learning outcomes: By the end of the session, the audience would be able to
- Describe the technology and processes used in the human bot collaboration framework to create Wikipedia content in regional languages.
- Appreciate the lessons from the pilot study that involved 50 computer science students generating 50K articles in 7 different domains.
- Explain the role of the volunteers to enrich the bot-generated articles.