AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Parsing as Pretraining

David Vilares, Michalina Strzyz, Anders Søgaard, Carlos Gómez-Rodríguez

AAAI 2020 · Architectures and efficiency › Attention and transformers · approach: Language model pretraining and induction · 38 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
Fine-tune BERT rather than freeze it (ablation)Parsing as Pretraining with BERT and a feed-forward label predictor · PTB testlabeled bracketing F1: 78.2 -> 93.5Improved
Fine-tune BERT rather than freeze it (ablation)Parsing as Pretraining with BERT and a feed-forward label predictor · EN-EWT UD testLAS: 63.0 -> 78.8Improved
Use BERT instead of ELMo with the feed-forward parser and fine-tune the encoder (ablation)Parsing as Pretraining · PTB testlabeled bracketing F1: 75.8 -> 93.5Improved
Use contextualized BERT vectors instead of random embeddings, with fine-tuning (ablation)Parsing as Pretraining · EN-EWT UD testLAS: 39.7 -> 78.8Improved
Add a 2-layer BiLSTM decoder (ablation)Parsing as Pretraining with frozen BERT · PTB testlabeled bracketing F1: 78.2 -> 92.2Improved
Add a 2-layer BiLSTM decoder (ablation)Parsing as Pretraining with fine-tuned BERT · PTB testlabeled bracketing F1: 93.5 -> 93.7No clear effect
Use pretrained word vectors instead of random embeddings (ablation)Frozen feed-forward parser · PTB testlabeled bracketing F1: 32.9 -> 41.4Improved
Use a task-specific BiLSTM decoder (ablation)Parsing as Pretraining with frozen input representations · PTB testlabeled bracketing F1: 78.2 -> 92.2; 41.4 -> 89.0Mixed

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