Ablation Studies in Artificial Neural Networks¶
Meyes, R., Lu, Melanie, de Puiseau, Waubert, C., Meisen, & Tobias. (2019). Ablation Studies in Artificial Neural Networks. arXiv.
Cited by¶
2 citations across 2 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Minimal Pairs
- Single-factor experiments are formal minimal-pair designs. Ablation in machine learning. Train two models differing only in whether they contain one component; attribute the performance change to that component.
This sourceMethods treatment of ablation as a tool to probe ANN internal representations by removing components and reading the change — minimal-pair experimentation in ML.
- Single-factor experiments are formal minimal-pair designs. Ablation in machine learning. Train two models differing only in whether they contain one component; attribute the performance change to that component.
Mechanisms¶
- Microdetail Ablation Suite
- Its strength is that it turns a rhetorical "that part doesn't matter" into a controlled result: after a good ablation suite you can say which details the pattern actually needs and which it can lose, which is exactly the knowledge downstream transfer requires (rebuild these, ignore those).
This sourceUses controlled ablations to identify which network parts carry distinct representations, which are important to classification, and which are redundant.
- Its strength is that it turns a rhetorical "that part doesn't matter" into a controlled result: after a good ablation suite you can say which details the pattern actually needs and which it can lose, which is exactly the knowledge downstream transfer requires (rebuild these, ignore those).
Verification¶
Does it exist? Not checked yet. This entry carries no identifier to resolve. It was extracted from the citation as written in the article, normalized, and deduplicated against the rest of the registry.
Does it back the claim? Not recorded. Neither this nor any other of the 2 citations of this work carries a recorded support check.
Support is checked per citation rather than per work — the same source can be cited soundly in one article and wrongly in another. Per-citation recording began recently, so a citation with no recorded check is a gap in the record rather than evidence it went unchecked.
See how references were verified.
Registry ID ref:47e6a2da69d5 · see in the full table