Module 3 - Epistemic Opacity in Machine Learning Models

Robert Michels: robert.michels@edu.ulisboa.pt

On this page, you will find the readings and an overview of the topics covered in the sessions of the module.

Session 1: Epistemological preliminaries and models in science.

Readings:

Session 2: Computer models, simulation, AI models in science; do they pose a special epistemic problem?

Readings:

Slides for session 2:
sem_epist_philsci-ulisboa_2024-02.pdf


Session 3

Readings:

  •  Explaining Machine Learning Decisions (Zerilli (2022)) - discussion of XAI (explainable AI), technical methods to make opaque ML models epistemically accessible to us -
Slides:
Rough summary of the Zirelli paper and our discussion:


Session 4

Readings:
Rough summary of the paper and our discussion:

Additional paper on AI in medicine:

Anexos