Computer Science
2 years ago

Revolutionizing Answer Selection: BERT Model Boosts Accuracy by Up to 18.7%

1 view
arXiv (Cornell University)
Utilizing Bidirectional Encoder Representations from Transformers for Answer Selection
Md Tahmid Rahman Laskar, Enamul Hoque, Jimmy Xiangji Huang

Paper Summary

The researchers used a pre-trained language model called BERT to improve answer selection in question answering tasks. By fine-tuning BERT on different datasets, they found significant improvements in both Question Answering and Community Question Answering tasks. The BERT model showed a maximum improvement of 13.1% in QA datasets and 18.7% in CQA datasets compared to previous methods.

Related papers

Theme background image
Computer Science
a year ago

BERT-Large Model Dominates Question Answering Tasks with 96% Accuracy

Natural Language Processing
Machine Translation
Visual Question Answering
1 view
Theme background image
Computer Science
a year ago

Revolutionizing Document Understanding: Transformer Models Boost Accuracy in Answering Questions

Visual Question Answering
Feature Retrieval
Transfer Learning
1 view
Theme background image
Computer Science
2 years ago

New Transformer Architecture Boosts Language Model Efficiency by 12 Units

Natural Language Processing
Machine Translation
Speech Recognition
1 view