Computational Competencies

Overview and goals

The Computational Competencies (Comp2) initiative is a strategic project at ETH Zurich aimed at integrating algorithmic modeling, data analysis, simulation, and programming into all Bachelor's and Master's programs. This initiative establishes "Learning Trajectories" (Lernspuren) that bridge foundational methods in computer science and mathematics with specific applications in various scientific disciplines. 

The strategic goals of the inititive are: 

• Curricular integration: all programmes provide tailored and coherent learning trajectories (Lernspuren) for Computational Competencies
• Foundations are introduced early and systematically and applied and deepened in disciplinary contexts.
• Computational competencies are not considered an add-on, they are an integral part of domain teaching
• Strengthen teaching: coordinate service courses, make Comp2 links explicit in domain lectures; use interactive notebooks and project?based learning
• Support teachers with didactic consultation, infrastructure and communities of practice.
• Build structures: common competence catalogue as a reference, technical & didactic support, and sustainable processes.

What this means: Students acquire practical competence in modelling & simulation, statistics & inference, machine learning, and programming ¡ª together with reflection on societal, legal and ethical implications. All this closely tied to and applied in their field of study.

ETH Competence Framework & Computational Competencies

Computational Competencies map onto the ETH Competence Framework. While most competencies are  in the block of subject specific competencies, Comp2 can be found in all areas of the ETH Competence Framework, as illustrated by the colors:

Generative AI & Competency Development

With generative AI, the value shifts from merely producing code to choosing, justifying, testing, and interpreting computational approaches. Effective tasks make thinking visible:
    ?    Disclose assumptions and discuss consequences for a given dataset/problem.
    ?    Justify method choices vs. at least one discarded alternative.
    ?    Explain given code and predict behaviour without execution.
    ?    Design test cases to expose weaknesses (incl. of AI?generated code).
    ?    Interpret results against domain expectations; document for reproducibility.
Catalogue note: GenAI remains embedded across ML, modelling, data handling, programming, and ethics ¡ª no separate ¡°AI competency¡± is defined in v0.42.

Design program?level learning trajectories

For sustainable impact, competencies should not remain isolated in individual courses. Programs define learning trajectories (Lernspuren) across semesters:
    ?    Introduce foundations early (often in service courses)
    ?    Apply and deepen repeatedly in disciplinary contexts
    ?    Ensure later courses build on shared tools and assumptions
The Curriculum Information Tool (CIT) visualises where competencies are taught and assessed ¨C making gaps and redundancies visible. Find out more!

Typical challenges
    ?    Long gaps without computational application
    ?    Over?reliance on single courses or individuals
    ?    Unclear handover between service and departmental courses