Visual dialog (VisDial) is a task which requires a dialog agent to answer a series of questions grounded in an image. Unlike in visual question answering (VQA), the series of questions should be able to capture a temporal context from a dialog history and utilizes visually-grounded information. Visual reference resolution is a problem that addresses these challenges, requiring the agent to resolve ambiguous references in a given question and to find the references in a given image. In this paper, we propose Dual Attention Networks (DAN) for visual reference resolution in VisDial. DAN consists of two kinds of attention modules, REFER and FIND. Specifically, REFER module learns latent relationships between a given question and a dialog history by employing a multi-head attention mechanism. FIND module takes image features and reference-aware representations (i.e., the output of REFER module) as input, and performs visual grounding via bottom-up attention mechanism. We qualitatively and quantitatively evaluate our model on VisDial v1.0 and v0.9 datasets, showing that DAN outperforms the previous state-of-the-art model by a significant margin.

Background COVID-19 caused significant confusion around the world, and dental education was no exception. Therefore, in line with the demands of the times, this study sought to determine the applicability of online active learning to dental education. Methods This study was conducted in the second semester of 2020 at a school of dentistry in a selective university in Korea. A total of 114 dental students were recruited. Participants were assigned to four different groups (lecture and discussion [LD], lecture and discussion with instructor’s worksheet [LW], self-study and discussion [SSD], and self-study and discussion with instructor’s worksheet [SW]) using the random breakout room function in the Zoom video conference application. Their final test scores were then analyzed using analysis of variance and the online active learning results were compared with the offline learning results. Results The scores were highest for the transfer type items in the SSD group, followed by the SW group and the two lecture groups, which had no significant differences. These scores and pattern differences between the groups were similar for all items. The results suggested that studying by oneself rather than simply listening to lectures enhanced the effects of the discussions and led to higher learning outcomes. In addition, the effect of the instructor's intervention in the middle of the discussion varied depending on the pre-learning activities of discussion. As with previous offline experiments, self-study followed by group discussion had higher learning outcomes for both the verbatim and transfer type items. Conclusions In agreement with the Interactive, Constructive, Active, and Passive (ICAP) framework and other active learning theories, the findings clearly indicated that online active learning was applicable to dental students, and when self-study precedes discussion, the learning is richer and the learning outcomes are better.

Background The ICAP framework based on Cognitive Science posits four modes of cognitive engagement: Interactive, Constructive, Active, and Passive. Focusing on the wider applicability of discussion as interactive engagement in medical education, we investigated the effect of discussion when self-study preceded it and further investigated the effect of generating questions before discussions. Methods This study was conducted in the second semester of 2018, and 129 students majoring in health professions, including medicine, dentistry, veterinary medicine, and nursing, participated. The students were assigned into four different trial groups, who were asked to fill out a Subjective Mental Effort Questionnaire after completing each session. Their performance in post-test scores and their mental efforts were analyzed. Results A Bonferroni test for group comparison indicated that the self-study and question-generated group had the highest performance and that the lecture and question-received group had the lowest performance when comparing the total score. By using a mediation model, it was confirmed that the participants who showed a higher level of testing mental effort also showed higher levels of studying and discussion mental effort. Conclusions Our findings support the ICAP framework and provide practical implications for medical education, representing the fact that students learn more when they are involved in active learning activities, such as self-study and question generation, prior to discussions.