A BERT Baseline for the Natural Questions

2019/01/24 by Chris Alberti, Kenton Lee, Alberti, Chris +3 · 16 citations
Computer Science · #Topic Modeling #Natural Language Processing Techniques #Multimodal Machine Learning Applications

paper · pdf · doi:10.48550/arxiv.1901.08634

Abstract

This technical note describes a new baseline for the Natural Questions. Our model is based on BERT and reduces the gap between the model F1 scores reported in the original dataset paper and the human upper bound by 30% and 50% relative for the long and short answer tasks respectively. This baseline has been submitted to the official NQ leaderboard at ai.google.com/research/NaturalQuestions. Code, preprocessed data and pretrained model are available at https://github.com/google-research/language/tree/master/language/question_answering/bert_joint.

Cited by

• No Answer is Better Than Wrong Answer: A Reflection Model for Document Level Machine Reading Comprehension
• A Discrete Hard EM Approach for Weakly Supervised Question Answering
• RikiNet: Reading Wikipedia Pages for Natural Question Answering
• Toward Deconfounding the Influence of Entity Demographics for Question Answering Accuracy
• Document Modeling with Graph Attention Networks for Multi-grained Machine Reading Comprehension
• Challenges in Information-Seeking QA: Unanswerable Questions and Paragraph Retrieval
• ActBERT: Learning Global-Local Video-Text Representations
• A question-answering system for aircraft pilots' documentation
• Pretrained Transformers for Text Ranking: BERT and Beyond
• Can NLI Models Verify QA Systems' Predictions?
• On Identifiability in Transformers
• Android Security using NLP Techniques: A Review
• ReCO: A Large Scale Chinese Reading Comprehension Dataset on Opinion
• VAULT: VAriable Unified Long Text Representation for Machine Reading Comprehension
• BERT-based knowledge extraction method of unstructured domain text
• Multi-passage BERT: A Globally Normalized BERT Model for Open-domain Question Answering

Related

• Frustratingly Easy Natural Question Answering
• The Death of Feature Engineering? BERT with Linguistic Features on SQuAD 2.0
• BERTSel: Answer Selection with Pre-trained Models
• Structured Pruning of a BERT-based Question Answering Model
• Automated Classification for Open-Ended Questions with BERT