Table 3 Types and difficulty level of automatic question generation and learner’s answer evaluation

Easy

Easy

Open-Cloze

Pino and Eskenazi (2009),

(Omit a word or phrase

(Match a word or phrase)

Question

Agarwal (2012),

from an informative

Das and Majumder (2017)

sentence)

Coniam (1997); Brown et al. (2005); Chen et al. (2006),

Moderate

Very Easy

Cloze Question

Hoshino and Nakagawa (2007); Pino et al. (2008),

(Main difficulty arise in

(Check option is true or

Agarwal and Mannem (2011); Correia et al. (2012),

distractors generation)

false)

Narendra et al. (2013)

Moderate

Very Easy

Multiple-Choice

Mitkov et al. (2006); Aldabe and Maritxalar (2010),

(Difficulty arise in

(Check option is true or

Question

Papasalouros et al. (2008); Bhatia et al. (2013),

distractors generation)

false)

Majumder and Saha (2014); Majumder and Saha (2015)

Difficult

Difficult

Subjective

Rozali et al. (2010); Dhokrat et al. (2012); Deena et al. (2020),

(Difficult in question

(Match learner’s answer

Question

Leacock and Chodorow (2003); Bin et al. (2008); Kakkonen et al. (2008),

formation, generally used

with handcraft model

Noorbehbahani and Kardan (2011); Dhokrat et al. (2012),

predefined template or

answer)

Islam and Hoque (2010); Ramachandran et al. (2015),

question pattern)

Sakaguchi et al. (2015)

Very Difficult

Very Difficult

Visual Question

Simoncelli and Olshausen (2001); Mora et al. (2016),

(Image tagging and

(Generate answer from

Mostafazadeh et al. (2016); Zhu et al. (2016); Yu et al. (2015),

question formation)

image-based question)

Ren et al. (2015); Zhang et al. (2017); Johnson et al. (2016); Jain et al. (2017)