Generalization without Systematicity¶
Lake, B. M., & Baroni, M. (2018). Generalization without Systematicity: On the Compositional Skills of Sequence-to-Sequence Recurrent Networks. ICML 2018.
Cited by¶
1 citation across 1 artifact.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Compositionality
- Connectionist models face the challenge of achieving compositional generalization: Lake-Baroni (2018) demonstrate with the SCAN benchmark that transformer language models, despite their empirical success, fail at systematic compositional generalization tasks.
This sourceIntroduces the SCAN benchmark; seq2seq RNNs succeed by mix-and-match but fail systematic compositional generalization (supports D22-230). NOTE: the paper tests RNN seq2seq models — the .md body at line 92 attributes the SCAN failure to 'transformer language models,' which is a factual error (see flag).
- Connectionist models face the challenge of achieving compositional generalization: Lake-Baroni (2018) demonstrate with the SCAN benchmark that transformer language models, despite their empirical success, fail at systematic compositional generalization tasks.
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:3413ff799872 · see in the full table