Recent proposed approaches have made promising progress in dialogue state tracking (DST). However, in multi-domain scenarios, ellipsis and reference are frequently adopted by users to express values that have been mentioned by slots from other domains. To handle these phenomena, we propose a Dialogue State Tracking with Slot Connections (DST-SC) model to explicitly consider slot correlations across different domains. Given a target slot, the slot connecting mechanism in DST-SC can infer its source slot and copy the source slot value directly, thus significantly reducing the difficulty of learning and reasoning. Experimental results verify the benefits of explicit slot connection modeling, and our model achieves state-of-the-art performance on Mul-tiWOZ 2.0 and MultiWOZ 2.1 datasets.

Objectives: During the coronavirus disease 2019 (COVID-19) self-quarantine period, the transition to online-course has profoundly changed the learning modes of millions of school-aged children and put them at an increased risk of asthenopia. Therefore, we aimed to determine associations of the total screen/online-course time with asthenopia prevalence among that children during the COVID-19 pandemic, and whether the associations were mediated by psychological stress.Methods: Asthenopia was defined according to a validated computer vision syndrome questionnaire (CVS-Q). We used CVS-Q to collect the frequency and intensity of 16 asthenopia-related eye symptoms of 25,781 children. Demographic features, eye care habits, visual disorders, lifestyle, psychological and environmental factors, were also collected.Results: The overall asthenopia prevalence was 12.1%, varying from 5.4 to 18.2% across grade/gender-classified subgroups. A 100-h increment of total screen/online-course time were associated with an increased risk of asthenopia by 9% [odds ratio (OR) = 1.09] and 11% (OR = 1.11), respectively. Mediation analysis showed that the proportions of total effects mediated by psychological stress were 23.5 and 38.1%, respectively. Age, female gender, having myopia or astigmatism, bad habits when watching screens were also risk factors. Conversely, keeping 34–65 cm between eyes and screen, increased rest time between classes, and increased eye exercise were all associated with a decreased risk.Conclusion: Our study indicated that the influence of long total screen or online-course time on psychological stress increases asthenopia risk. The findings of this study have provided a new avenue for intervening screen-related asthenopia in addition to incorporating a reasonable schedule of online courses into educational policy.