Collective intelligence for Sweden´s universities

Important results from the project

Kollin is building a platform to enable collective intelligence among university students around the country. The project´s objective was to lay its foundation, through smart systems and a strong community. We have, since the start of the project, achieved important milestones in both technical and commercial areas, e.g. through the setup of an architecture for distributed generation, curation and formatting of academic content. We have tested and validated clustering and NLP models for content indexing, quadrupled Q&A among students and increased users by 75%.

Expected long term effects

Positive results on market penetration, product development and the degree of automation of our processes have been achieved. With regards to AI, both supervised and unsupervised models were developed. A lot of focus was put on the latter type, as it enables greater upscaling and is a less researched area. The innovation there belongs to the automatic and relevant naming of clusters with small documents, using open-source content resources. More research is planned so that it can be used commercially. In the meantime, the supervised model could be put into production.

Approach and implementation

A high degree of synchronization between the development of smart systems and the community were key factors for success. On the more technical front, relevant AI methods and possible back-end tools needed to be thoroughly explored, implemented, evaluated, dismissed or integrated. to achieve a higher utilization rate of the students. User tests were performed continuously during the project, both on prototypes of the website and when these were implemented in production, with the aim of maximizing user engagement.

External links

Kollins hemsida