SDIS-researchers Belal Abu Naim and Yasin Ghafourian coached students of the WU Vienna. Together they explored research questions of the project AIMS5.0.
In course of their “Data Science Lab” students at the WU Vienna gained hands on experience in project work with the help of SDIS-researchers Belal Abu Naim and Yasin Ghafourian. Supervisors from different industries guided the students through some tasks connecting to real-life projects.
A group of students decided to work on the AIMS5.0 project and worked through it throughout the semester with help of the SDIS-researchers. Use Case 11 of the EU-funded project is developing an autonomous framework to control a vertical farm. A key component of this framework is the NLP pipeline, which is responsible for extracting knowledge from agricultural literature—like research papers and books—related to specific crops.
For example, in the case of growing salads, the pipeline should be able to process the literature and summarize it into a set of cultivation rules, such as:
“If the plant is in this growth phase, the temperature should be between X and Y, the lighting should be Z, etc.” The idea is that this pipeline will be flexible—so the researchers can simply change the input literature and adapt it to other crops as needed.
The students’ task was to experiment with different large language models to see which one can most accurately extract these kinds of rules from PDFs. The output should ideally be a condensed, structured set of guidelines (e.g., 50 summarized lines) that reflect the cultivation guide.
Now at the end of the semesters the students presented their results.