AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

Pretrained Encyclopedia: Weakly Supervised Knowledge-Pretrained Language Model

Wenhan Xiong, Jingfei Du, William Yang Wang, Veselin Stoyanov

ICLR 2020 · Vision-language and multimodal › CLIP and zero-shot · approach: Language models for zero-shot reasoning · 25 citations (OpenAlex)

Read the paper

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Add type-constrained entity replacement pretrainingBERT-base · Zero-shot fact completion (10 relations)Hits@10: 16.3 -> 28.9Improved
Add entity replacement pretrainingOur BERT · SQuAD devF1: 90.5 -> 91.3Improved
Add entity replacement pretrainingOur BERT · WebQuestionsF1: 35.5 -> 37.9Improved
Add entity replacement pretrainingOur BERT · TriviaQAF1: 53.2 -> 56.7Improved
Add entity replacement pretrainingOur BERT · Quasar-TF1: 46.1 -> 49.9Improved
Add entity replacement pretrainingOur BERT · SearchQAF1: 61.9 -> 63.3Improved
Add entity replacement pretrainingOur BERT · FIGERAcc: 54.53 -> 60.21Improved
Add paragraph ranking scoresWKLM · TriviaQAF1: 56.7 -> 63.1Improved
Add paragraph ranking scoresWKLM · Quasar-TF1: 49.9 -> 52.2Improved
Add paragraph ranking scoresWKLM · WebQuestionsF1: 37.9 -> 41.8Improved
Remove masked language modeling objective (ablation)WKLM · SQuADF1: 91.3 -> 87.6Hurt
Remove masked language modeling objective (ablation)WKLM · TriviaQAF1: 56.7 -> 52.5Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy