Portions of this blog post are excerpted from the accompanying podcast episode and from notes shared between the SciStarter team and Kobi Gal’s research team.
Smart Project Recommendations on SciStarter
With thousands of projects listed on SciStarter, a main challenge can be finding the right project, one that really suits your needs and your interests. After meeting at a workshop on the Open Science of Learning hosting by CRI, Kobi Gal, a leading expert in human-centered artificial intelligence, and Darlene Cavalier, the founder of SciStarter, collaborated (with support from NESTA, a UK-based innovation foundation) to create a smart recommendation system to help SciStarter users find the right project.
In a new podcast episode, Kobi and Na’ama Dayan, a graduate student at Ben Gurion University and a member of Kobi Gal’s research team, chat with Caroline Nickerson from the SciStarter team about the new system and how YOU can help us test it over the next few weeks. A transcript is available here.
The goal of the new system is to personalize project recommendations on SciStarter, so participants can easily discover the projects most suitable for them. The team hopes that personalized Artificial Intelligence (AI)-powered recommendations can generate improved scientific and learning outcomes and increase the satisfaction of volunteers.
The system matches users with projects selected by other users with similar characteristics, based on their profiles and activities. This matching is done through grouping users into five anonymous cohorts based on their SciStarter contributions (types of projects contributed to, number of projects contributed to, frequency of contributions, time spent doing the project, and more). Users can access these recommendations when logged in, both on the SciStarter homepage and as a sidebar on project pages.
As Kobi says in the podcast, “AI is a collection of techniques and technologies that compliment human abilities, rather than replace human abilities, and help us to do better jobs, whether as teachers, judges, doctors, or computer users.”
Kobi explains in the episode why some people may be afraid of AI, elaborating about how this team combats these fears with intentional design. “AI, just like any tool, like any technology, can be used whether for the good or for the bad…the research we do here in SciStarter, our collaboration, is aimed at mitigating these fears in that we create a system that is able to help people achieve better satisfaction, find the projects that match their interests, rather than trying to tell them what to do, or to manipulate the way they interact on the site.”
A user can opt out at any time. If a user opts out, they will get the default, fixed list of popular projects on their SciStarter homepage and on project page sidebars.
How it works
Recommendation systems (commonly used in e-commerce, in the news, and on social media sites) use algorithms that analyze past behavior to recommend items to users, relying on hundreds of thousands of data instances. As Na’ama explains in the podcast, an “algorithm is like a set of steps in order to achieve a goal. It’s like following the instructions of a recipe to make a cake.”
The system matches users with items that are liked by similar users, with similarity between users based on their past behavior.
Na’ama says in the podcast interview that the team uses five different algorithms. Projects can be recommended based on project similarity, meaning that “a user is recommended projects that are similar to each other, in the sense that they’re are both done online or that similar users have participated in this project,” or projects can be recommended based on user similarity, meaning that, “if John participates in CoCoRaHs and John and I are similar in the sense of our past activities…it is reasonable that I would also like to participate in CoCoRaHS, because John and I are similar.”
For users with significant history with the SciStarter platform, the recommendation engine will make predictions based on the algorithms. For users that are new to the SciStarter platform, the engine will rely on recommending most popular projects to users.
According to a report from the National Academies of Sciences, Engineering, and Medicine, citizen scientists’ motivations are “strongly affected by personal interests,” and participants who engage in citizen science over a long period of time “have successive opportunities to broaden and deepen their involvement.” Thus, it seems that sustained engagement through the use of intelligent recommendations can improve data quality and scientific outcomes for the projects and the public.
All data collected and analyzed during this experiment on SciStarter will be anonymized. This includes all project participation data and clickstream data. The team will analyze the data, comparing the cohorts to examine which types of users have contributed to different projects in terms of number of visits, saves, length of time spent in the project, and frequency of contributions. The team will also conduct surveys of SciStarter users and seek community feedback (yes, they want to hear from you!).
How will we know this worked? Kobi told us in the podcast, “If we did things right, then these new recommendations will actually make people better contributors to SciStarter.”
The team will share their output with the project owners and the volunteering community, providing them with insights on other types of projects their participants engaged with. If the project is successful, the study will yield a reproducible algorithm that citizen science platforms can use for intelligent recommendations, as well as a generalized approach for improving collective intelligence in citizen science by connecting users, data, and AI.
So what does this mean for you? You can head over to SciStarter right now to test the “friendly parrot on the shoulder,” as Kobi calls the recommendations. Did we nail it? Are you able to more easily find the right project for you? We’d love to hear your opinion on this, and you can let us know what you think via email at firstname.lastname@example.org.
Thank you to Kobi Gal, Na’ama Dayan, Avi Segal, Zhixing, and others for their work on this project.