BWhat changed
Pretrained a sequence-to-sequence transformer by corrupting text and reconstructing it, reporting strong generation and comprehension transfer.
Scientific scope
Pretrained a sequence-to-sequence transformer by corrupting text and reconstructing it, reporting strong generation and comprehension transfer. Scope limitation: The 2019 preprint was later published at ACL 2020; the paper does not invent denoising autoencoders or transformer encoder-decoders.
Evidence
The retained primary-source abstract, bibliographic record and stated review scope support this provisional assessment. No independent experimental replication by WRS is claimed. The 2019 preprint was later published at ACL 2020; the paper does not invent denoising autoencoders or transformer encoder-decoders.
Contributors
- The named authors listed in the linked research record · SCIENCE: Complete source byline retained in the research record; human research direction, method development and evaluation. Individual labor shares are not measured.
- Lewis and colleagues: BART denoising seq2seq pre-training — original paper
- Lewis and colleagues: BART denoising seq2seq pre-training — official research page
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction