15. ESAC. Shaping the future of Science Operations through Natural Language Processing - ESAC Trainees
15. Shaping the future of Science Operations through Natural Language Processing.
ESA supervisor: Sandor Kruk
Collaborator(s): Catarina Alves de Oliveira, Nora Luetzgendorf, Jan Reerink, Pablo Gomez
One of the areas critical to the success of ESA's science missions is science operations. How can we simplify the process for upcoming missions to effectively express their science operation requirements while capitalizing on commonalities across multiple missions?
The primary objective of this traineeship is to harness the latest advancements in natural language processing and implement these techniques within the realm of space science missions at ESA. In this project, you will employ open-source, large language models like LLAMA2 and enhance their capabilities through fine-tuning using information sourced from ESA's science documents. The ultimate goal is to construct an AI agent capable of extracting knowledge, summarizing information, and identifying similarities within a substantial body of text pertaining to scientific requirements. This agent will be employed by new missions to tap into the pre-existing requirements upon which science operation centers have been established, reducing the need to create entirely new requirements solely for the specific operations of new missions. This approach will facilitate a more streamlined and faster establishment of science operation requirements.
Project duration: 6 months.
Desirable expertise or programming language:
- An interest in natural language processing and a basic understanding of its principles.
- Programming experience in Python.
- A background or coursework in computer science, data science, or a related field is a plus.
To apply for this project please fill in an online application form through the following link.