Towards the application of calibrated Transformers to the unsupervised estimation of question difficulty from text

Aims To investigate the motivations for, and perceived benefits of, undertaking senior house officer (SHO) posts, and to explore the career pathways of those who do, examining trends in successive cohorts. Method Postal cross-sectional questionnaire survey of all dental and maxillofacial SHOs (DF2s) who had worked for two South London hospitals within the previous nine years (n=137). Respondents were grouped into three cohorts to enable responses to be examined in relation to respondents’ entry to their first SHO post. Results There were responses from 83 (61%) potential participants. The most frequent motivation for carrying out SHO posts from 79 (95%) of the respondents was the desire to ‘learn from experienced clinicians’. The most common perceived benefit reported by those who had completed posts at the time of the survey was ‘an improved understanding of the role of the hospital dental service’ from 68 (97%) of those who answered this question. Difficulty in securing a job in general dental practice was not reported as a notable motivating factor, either before or after the implementation of the new dental contract. ‘Fulfil approved training post requirements for postgraduate examinations’ reduced as a motivator from 28 (88%) for the earlier cohort of SHOs to nine (36%) for the more recent cohort. Fifty-four of 78 (69%) respondents declared a definite plan to seek admission to the General Dental Council Specialist Lists in future, 24 (83%) in the first cohort, compared with 11 (46%) in the last ( P=0.05). Of the males, 13 (52%) were significantly more likely to report that they were currently working in general practice compared with 15 (27%) females ( P=0.028). Conclusion The findings suggest that multiple benefits are identified from undertaking SHO posts. However, some of the motivations for undertaking SHO posts may have changed over the nine-year period investigated. Possible influences are discussed. This paper highlights the perceived benefits of junior training posts at a time of significant transition within the profession.

Download Full-text

Question Difficulty Priori Evaluation Based on Fuzzy Logic in Programming System

2021 IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA) ◽

10.1109/icaica52286.2021.9498258 ◽

2021 ◽

Author(s):

Qing Du ◽

Wei Zheng ◽

Chuanlin Xia

Keyword(s):

Fuzzy Logic ◽

Programming System ◽

Question Difficulty

Download Full-text

Unsupervised Learning of Question Difficulty Levels Using Assessment Responses

Computational Science and Its Applications – ICCSA 2017 - Lecture Notes in Computer Science ◽

10.1007/978-3-319-62392-4_39 ◽

2017 ◽

pp. 543-552

Author(s):

Sankaran Narayanan ◽

Vamsi Sai Kommuri ◽

N. Sethu Subramanian ◽

Kamal Bijlani ◽

Nandu C. Nair

Keyword(s):

Unsupervised Learning ◽

Question Difficulty

Download Full-text

Intelligent Questioning System Based on Fuzzy Logic

Research Methods ◽

10.4018/978-1-4666-7456-1.ch029 ◽

2015 ◽

pp. 620-642

Author(s):

Omer Deperlioglu ◽

Guray Sonugur ◽

Kadir Suzme

Keyword(s):

Fuzzy Logic ◽

Nearest Neighbor ◽

Difficulty Level ◽

Knowledge Level ◽

Clustering Methods ◽

Multiple Choice Questions ◽

Student Knowledge ◽

Question Difficulty ◽

Knowledge Levels ◽

And Performance

One of the most important functions of distance learning systems is determining the student knowledge level and performance clearly. In traditional education systems, students can be assessed in single-stage via tests and homework studies, which consist of multiple-choice questions. However, this method cannot provide accurate results since it is not able to evaluate student knowledge level and question difficulty level. In this chapter, a system and software structure that can determine student knowledge levels, topic difficulty level, and question difficulty levels according to instant student answers for the exam is introduced. In forming student knowledge levels, content monitoring and test data taken from distance education vocational school were used. In this way, more accurate results have been obtained. The fuzzy logic technique has been used to determine (classify) student knowledge levels and topic difficulty levels clearly. In order to determine next questions adaptively, the stored questions have been classified with division clustering methods, and the most suitable questions for the related student knowledge level have been found by using the nearest neighbor algorithm.

Download Full-text