| Abstrakt: |
2.1.) I would like to put the focus here on the clustering algorithms, and on what is arguably a component of EC, namely that EC*: Clustering algorithms identify natural kinds. Unlike supervised algorithms, clustering algorithms do not come with clear success criteria and accompanying theoretical success guarantees. The domain of unsupervised learning, unlike supervised learning and reinforcement learning, has largely been neglected in philosophical work on machine learning. [Extracted from the article] |
Full text from SpringerLink