Generating and Evaluating Explanations of Attended and Error-Inducing Input Regions for VQA Models

Attention maps, a popular heatmap-based explanation method for Visual Question Answering (VQA), are supposed to help users understand the model by highlighting portions of the image/question used by the model to infer answers. However, we see that users are often misled by current attention map visualizations that point to relevant regions despite the model producing an incorrect answer. Hence, we propose Error Maps that clarify the error by highlighting image regions where the model is prone to err. Error maps can indicate when a correctly attended region may be processed incorrectly leading to an incorrect answer, and hence, improve users’ understanding of those cases. To evaluate our new explanations, we further introduce a metric that simulates users’ interpretation of explanations to evaluate their potential helpfulness to understand model correctness. We finally conduct user studies to see that our new explanations help users understand model correctness better than baselines by an expected 30% and that our proxy helpfulness metrics correlate strongly (rho>0.97) with how well users can predict model correctness.

Should students change their answers on multiple choice questions?

Multiple choice exams are ubiquitous, but advice on test-taking strategies varies and is not always well informed by research. This study evaluated the question of whether students benefit or are harmed when they change their initial answers on multiple choice questions in the context of physiology and biology courses. Previously marked examinations were reviewed for eraser marks that indicated answer changes, and the impact of these changes on exam grades was tabulated. In addition, faculty and students were surveyed for their opinions about changing answers. A plurality of faculty (36%) reported a belief that answer changes usually harm student grades, whereas a slim majority of students (51%) believed that answer changing helped their scores (χ2 = 60.52, P < 0.0001). Empirically, across two exams, students changed their answer from an incorrect answer to a correct one 2.8 times (SD 2.2) compared with 1.0 time (SD 1.4) changing in the negative direction. Therefore, on average, students benefited ( V = 123.5, P < 0.0001) from answer changing. Furthermore, comparing across two exams in the same course, some students were consistently more likely to change their answers than others (adjusted R2= 0.23, P < 0.0001), but the impact of changing answers on the first exam provided no prediction of how much a student would benefit from answer changing on the second exam (adjusted R2= −0.004, P = 0.42). These data support the argument that students should be advised to review and revise responses to exam questions before submitting them.

Anchoring effect of performance feedback on accuracy of metacognitive monitoring in preschool children

Preschool children are generally inaccurate at evaluating past and predicting future performance. The present study examines the effect of performance feedback on the accuracy of preschoolers’ predictive judgments and tests whether performance feedback acts as an anchor for postdictive judgments. In Experiment 1, preschool children (n = 40) solved number patterns, and in Experiment 2 they solved object patterns (n = 59). The results in both experiments revealed, firstly, that children receiving performance feedback made more accurate predictive judgments and lowered their certainty after their incorrect answer. Secondly, the children relied on performance feedback more than on actual task experience when making postdictive judgments, indicating that performance feedback was used as an anchor for subsequent postdictive judgments.

Knowledge of CT exposure parameters among Norwegian student radiographers

Background Improvements in the competency levels of student radiographers in computed tomography examinations (CT) are important due to the increasing number of these examinations being undertaken in imaging departments. The present study assesses the knowledge of student radiographers regarding CT exposure parameters. Methods The level of knowledge related to CT exposure parameters was evaluated using a twenty-one-item questionnaire that was distributed to final-year student radiographers. The questionnaire consisted of questions around CT exposure parameters and either allowed respondents to answer “true,” or “false” or choose a response from a range of responses where only one answer was correct. Correct answers were given one mark, while no mark was given for an incorrect answer. The score out of possible 21 was converted to a percentage, with a higher percentage signifying greater knowledge. Results Seventy-two students completed and returned the questionnaire, resulting in a 71% response rate. The mean score was 53%. Only 33% of students correctly identified that kilovoltage peak (kVp) should be increased when patients have metallic implants, and milliampere seconds (mAs) should be increased as body part thickness increases. No one answered all the questions correctly. There was no significant knowledge difference between students who had CT facilities on campus and those that did not. Conclusion Overall, student radiographers’ knowledge of CT exposure parameters was reported to be satisfactory.

A ANALYZING THE LEVEL OF THE STUDENTS’ READINGCOMPREHENSION IN COMPREHENDING NARRATIVETEXT

The research aimed at finding out the analyzing of students’ reading comprehension at the second year of SMU Negeri 4 Bulukumba Kab. Bulukumba. The findings of this research that the researcher known the ability of the students’ reading comprehension on level at the second year of SMA Negeri 4 Bulukumba to have “Low” level. Based on the finding indicated that the second year of students’ SMA Negeri 4 Bulukumba in answer or identifying reading comprehension test, there were 60% students got correct answer in ( answer with literal) and 40% students had incorrect answer. Then 40% students had correct answer in (Interpretation) and 60% students had incorrect answer. Then 30% students had correct answer in (Critical) and 70% students had incorrect answer. Then 35% students had correct answer in (Creative) and 65% students had incorrect answer. Based on the finding about the students competence on reading comprehension level, some problems or difficulties were faced by the second grade students of SMA Negeri 4 Bulukumba in analyzing reading test with correct answer which were answer with literal, interpretation, critical and creative. Based on the result of questionnaire of students, the researcher finds out some solution from the problem of the students in reading. The students are suggested to practice reading anything that they like, for example reading a article, reading, essay, reading a book, reading newspaper and reading novel every day. The students practice interpreting what has been reading in improving reading comprehension. Keywords: Student, Reading Comprehension, Level, Analyzing

Analysis of Students’ Vocabulary in Learning English

The purpose of this study was to analysis students’ vocabulary in learning English of Ninth grade at SMP Swasta Yayasan Pangeran Antasari, Medan in the academic year of 2018/2019. The researchers used descriptive qualitative method to describe and implementing factual condition and characteristic. There are 72 students that decided two classes at ninth grade as the population. But the researcher only took one class consists of 36 students as the sample. To collecting data, researchers use questionnaire and vocabulary comprehension test as procedure. In analyzing the data, the researchers identifying the difficulties of students’ vocabulary and cause difficulties by analyzing students answer sheets and their mistakes. There were 25 students (70%) incorrect answer in the questions. There were 6 students (16%) in modal, there were 8 students (23%) in simple past tense, There were 8 students ( 24 %) in simple present tense, 3 students (7%) troubled in simple continuous tense. Meaning. Students who gave incorrect answer in the questions were 29 students (82%). Others difficulties is translation being other types of the questions that examined in this research. it was found that there were 23 students (64%) who were unable to give the correct answer. Concluding the cause difficulties of students’ vocabulary in learning English are the first students difficulties in pronouncing the words. Grammatical form be the second cause because the written form is different from the spoken form in english. The last, students does not give their attention to study english more, knowing a words much more, and limitations their time to sources of information about new words vocabulary. The last, forgot the words easily. Keywords : Analysis Students’ Vocabulary, Difficulties, Cause

EXAMINING GENDER AND NATIONALITY BIAS IN DECISION-MAKING BY ENGINEERING STUDENT TEAMS

This study uses two-stage team quizzes to assess differences in team decision-making based on the factors gender and nationality. Over 200 teams in two different engineering design courses delivered using Team-Based Learning across five years were considered. In the two-stage quizzes, individuals first committed to their own answers, and then the team discussed the same questions and answered as a group. Cases where an individual was incorrect and the team adopted that same incorrect answer were used as a measure of influence of that individual on team decision-making (i.e., “pushing” behaviour by the individual). Similarly, cases where an individual was correct but the team adopted a different (incorrect) answer were used as a measure of lack of influence (i.e., “switching” behaviour by the individual). Overall, no significant gender or nationality differences were found in pushing behaviours. Male students and international students were found to be more likely to engage in switching behaviours. The overall differences in switching were modest (0.3-0.4% difference per question), but this translates to between 5 and 15 more male/international students engaging in switching behaviours in a typical 75- to 150-student course.

What Went Wrong for Bad Solvers during Thematic Map Analysis? Lessons Learned from an Eye-Tracking Study

Thematic map analysis is a complex and challenging task that might result in map user failure for many reasons. In the study reported here, we wanted to search for differences between successful and unsuccessful map users, focusing—unlike many similar studies—on strategies applied by users who give incorrect answers. In the eye-tracking study, followed by a questionnaire survey, we collected data from 39 participants. The eye-tracking data were analyzed both qualitatively and quantitatively to compare participants’ strategies from various perspectives. Unlike the results of some other studies, it turned out that unsuccessful participants show some similarities that are consistent across most analyzed tasks. The main issues that characterize bad solvers relate to improper use of the thematic legend, the inability to focus on relevant map layout elements, as well as on adequate map content. Moreover, they differed in the general problem-solving approach used as they, for example, tended to choose fast, less cautious, strategies. Based on the collected results, we developed tips that could help prevent unsuccessful participants ending with an incorrect answer and therefore be beneficial in map use education.

ANALYSIS ON THE STUDENTS’ COMMON DIFFICULTIES IN TOEFL READING COMPREHENSION (A CASE STUDY ON THE SECOND SEMESTER OF ENGLISH STUDY PROGRAM AT ISLAMIC UNIVERSITY OF INDRAGIRI TEMBILAHAN)

The problem in this study was the students’ common difficulties in TOEFL Reading Comprehension. This study only used one variable. The research method used in this study was a descriptive research. The purpose of this study was to know the students’ common difficulties in TOEFL Reading Comprehension on the second semester of English Study Program at Islamic University of Indragiri Tembilahan. The total of the sample used in this study were 23 students, the technique was used to take the sample is total sampling. In collecting the data, the researcher used a test. After the data has been collected, the result of test showed that the skills that are considered to be the most difficult ones in TOEFL reading, which is skill 3. This skill is identifying stated detail questions correctly. It is the highest percentage (84%) with the total of incorrect answer is 210 from 5 items. Then, it is followed by skill 5 (82%) with the total of incorrect answer is 82 from 2 items, which is finding pronoun reference. It is the second most difficult skill for the students. The third most difficult skill is skill 7. It is responding of transition question achieve 81% with the total of incorrect answers is 81 from 2 items. Next, skill 10 is considered to be the fourth most difficult skill in TOEFL reading comprehension. The questions are about the use of context to give meaning of difficult word which is counted 80 % from the total of incorrect answers is 120 from 3 items. Then, the fifth most difficult skill in TOEFL reading comprehension is skill 1, which is consisting the questions about answering main idea questions correctly, students made 78% of the total incorrect answers is 156 from 4 items.