j d karpicke et al 2012 r j hift 2014 d rohrer et al 2010
play

[ J. D. Karpicke et. al., 2012 ] [ R. J. Hift. 2014 ] [ D. Rohrer - PowerPoint PPT Presentation

[ J. D. Karpicke et. al., 2012 ] [ R. J. Hift. 2014 ] [ D. Rohrer et. al., 2010 ] ? QG-Net: Data-Driven Question Generation Model for Educational Content Jack Wang June 25, 2018 Generated Input text QG-Net question one of the


  1. [ J. D. Karpicke et. al., 2012 ] [ R. J. Hift. 2014 ] [ D. Rohrer et. al., 2010 ]

  2. ?

  3. QG-Net: Data-Driven Question Generation Model for Educational Content Jack Wang June 25, 2018

  4. Generated Input text QG-Net question … one of the positive What is one of the positive contributions of deviance is contributions of deviance? that it fosters social change . ✓ Fluent ✓ relevant

  5. Generated Input text QG-Net question ✓ Generates fluent and relevant questions ✓ Adaptive to texts from various subjects

  6. ✓ Fluent and relevant questions QG-Net ✓ Adaptive to texts from various subjects LSTM LSTM network network Reader Generator 0100010101 Attention 0100110100 mechanism 1010001010 0111010101 Copy 01010101... mechanism Intermediate representation What is one of the positive … one of the positive contributions of deviance is contributions of deviance? that it fosters social change .

  7. Experiment overview ● QG-Net learns to generate questions using SQuAD [ ] Source sentence Answer Generated (from Wikipedia) (selected from text) question Nikola Tesla was a What is Tesla’s Serbian American Serbian American nationality? inventor, …

  8. Experiment overview ● QG-Net learns to generate questions using SQuAD ● We test it on three OpenStax textbooks ○ Pre-processed the textbook texts to the same format as SQuAD 1. M. Heilman, 2010 ○ Compared to 2 of the best existing models (baseline1 1 , baseline 2 2 ) 2. X. Du et. al., 2017

  9. Qualitative evaluation result #1 (Sociology) Weber noted that different groups were affected differently based on education, race and gender , and that people's’ reactions to inequality were moderated by class differences and rates of social mobility, ... Chapter 1.3 QG-Net : Along with education and race, what did Weber believe different groups were affected by? Baseline 1: What was education, race, and,? Baseline 2: Along with education, race, race, and race, what other groups were affected by the Kinsey scale? Chapter 12.1

  10. Qualitative evaluation result #2 (History) The 1830 Indian Removal Act and subsequent displacement of the Creek, Choctaw, Chickasaw, Seminole, and Cherokee tribes of the southeast fulfilled the vision of a white nation and became one of the identifying characteristics of the Chapter 10.4 age of Jackson. QG-Net : What act became one of the identifying characteristics of the age of Jackson? Baseline 1: What did and subsequent displacement of the Creek, Choctaw, Chickasaw, Seminole, and Cherokee tribes of the southeast fulfil the vision of a white nation? Baseline 2: What was the name of the act that caused a white nation to become a white nation?

  11. Qualitative evaluation result #2 (History) The 1830 Indian Removal Act and subsequent displacement of the Creek, Choctaw, Chickasaw, Seminole, and Cherokee tribes of the southeast fulfilled the vision of a white nation and became one of the identifying characteristics of the age of Jackson. QG-Net : What act became one of the identifying characteristics of the age of QG-Net: Jackson? ✓ Generates fluent and relevant questions ✓ Is adaptive to text from various subjects Baseline 1: What did and subsequent displacement of the Creek, Choctaw, Chickasaw, Seminole, and Cherokee tribes of the southeast fulfil the vision of a white nation? Baseline 2: What was the name of the act that caused a white nation to become a white nation?

  12. Human evaluation experiment ● We sample 150 sentences from each textbook ● We use the three models to generate questions ● Human evaluators evaluate the quality of generated questions ○ Fluency ○ relevance ● Human evaluators pick which question(s) could have been generated by human ○ Can choose none of them

  13. Human evaluation results QG-Net: ✓ Generates superior questions in terms of Baseline 1 Baseline 2 both fluency and relevance ✓ Works well for inputs from various subjects

  14. Human evaluation results Perceptually, QG-Net generates questions of Baseline 1 superior quality than those generated by Baseline 2 baseline models.

  15. Acknowledgements Rich Baraniuk Andrew Lan Phil Grimaldi Drew Waters Weili Nie John and Ann Doerr Arthur & Carlyse Ciocca Charitable Foundation

  16. Questions? Generated Input text QG-Net question QG-Net is on Github! https://github.com/moonlightlane/QG-Net jzwang@rice.edu

Recommend


More recommend