2025-07-10 | DEST Hackathon (Potsdam, Germany)
July 10–11
This is the signup form for the first DEST hackathon in Potsdam on the Telegrafenberg. The event is organised by the Evidence for Climate Solutions group at PIK as part of the Wellcome-funded DESTinY project.
Everyone who is interested in AI for digital evidence synthesis / technology assisted reviews (text mining to summarise findings published in research articles) is warmly welcome! The main themes for the workshop will be user interfaces and evaluation of AI-assisted methods to automate ( parts of) the synthesis pipeline.
Example projects could be to compile a list of existing (interactive) platforms to share syntheses and to collect useful features that we should be able to replicate. At the same time, we might also want to test some of our early demonstrator products to measure real-world benefits over conventional approaches.
This is not just for people with programming skills; anyone is needed and warmly welcome. Some experience in evidence synthesis (meta-studies, systematic maps, systematic reviews) is helpful. Data scientists get the opportunity to work with large human-annotated datasets to evaluate their tools or experiment with ideas such as preparing (semi-)supervised fine-tuning data for domain-specific LLMs.
The hackathon is co-hosted with the DESTinY and ICASR meeting. If you already indicated your hackathon participation there, you don’t need to sign up here too.
Project ideas
We are still working on preparing topics and we won’t limit the choices by our proposals. This list is to give you an idea of potential projects and projects where someone committed to prepare input.
- Towards acceptable minimum quality requirements for AI use in evidence synthesis; using results from our large community survey.
- Effective user interfaces: From a collection of interactive web interfaces, we want to run usability studies to find out what actually improves workflows. We’d also like to sketch out visions for next-generation data sharing platforms and tools for evidence synthesis.
- Stopping methods: Priority screening has the potential to save a lot of effort by seeing (potentially) relevant records first and then stopping early. Following the ICASR meeting, we’d like to invite others to contribute to our benchmarking framework.
- Enhancement processes: Play with the new DESTinY repository, to learn how to use the interchangeable formats that we will base our future developments on. This helps us to collect feedback and you could bring your favourite tool into the DESTinY ecosystem.