Different texts shall by nature correspond to different number of keyphrases. This desideratum is largely missing from existing neural keyphrase generation models. In this study, we address this problem from both modeling and evaluation perspectives.We first propose a recurrent generative model that generates multiple keyphrases as delimiter-separated sequences. Generation diversity is further enhanced with two novel techniques by manipulating decoder hidden states. In contrast to previous approaches, our model is capable of generating diverse keyphrases and controlling number of outputs.We further propose two evaluation metrics tailored towards the variable-number generation. We also introduce a new dataset (ST A C KEX) that expands beyond the only existing genre (i.e., academic writing) in keyphrase generation tasks. With both previous and new evaluation metrics, our model outperforms strong baselines on all datasets.

Background Continued use of mHealth apps can achieve better effects in health management. Gamification is an important factor in promoting users’ intention to continue using mHealth apps. Past research has rarely explored the factors underlying the continued use of mobile health (mHealth) apps and gamification’s impact mechanism or path on continued use. Objective This study aimed to explore the factors influencing mHealth app users’ intention to continue using mHealth apps and the impact mechanism and path of users’ feelings induced by gamification on continued mHealth app use. Methods First, based on the expectation confirmation model of information system continuance, we built a theoretical model for continued use of mHealth apps based on users’ feelings toward gamification. We used self-determination theory to analyze gamification’s impact on user perceptions and set the resulting feelings (competence, autonomy, and relatedness) as constructs in the model. Second, we used the survey method to validate the research model, and we used partial least squares to analyze the data. Results A total of 2988 responses were collected from mHealth app users, and 307 responses were included in the structural equation model after passing the acceptance criteria. The intrinsic motivation for using mHealth apps is significantly affected by autonomy (β=.312; P<.001), competence (β=.346; P<.001), and relatedness (β=.165; P=.004) induced by gamification. The intrinsic motivation for using mHealth apps has a significant impact on satisfaction (β=.311, P<.001) and continuance intention (β=.142; P=.045); furthermore, satisfaction impacts continuance intention significantly (β=.415; P<.001). Confirmation has a significant impact on perceived usefulness (β=.859; P<.001) and satisfaction (β=.391; P<.001), and perceived usefulness has a significant impact on satisfaction (β=.269; P<.001) and continuance intention (β=.273; P=.001). The mediating effect analysis showed that in the impact path of the intrinsic motivation for using the mHealth apps on continuance intention, satisfaction plays a partial mediating role (β=.129; P<.001), with a variance accounted for of 0.466. Conclusions This study explored the impact path of users’ feelings induced by gamification on the intention of continued mHealth app use. We confirmed that perceived usefulness, confirmation, and satisfaction in the classical continued use theory for nonmedical information systems positively affect continuance intention. We also found that the path and mechanism of users' feelings regarding autonomy, competence, and relatedness generated during interactions with different gamification elements promote the continued use of mHealth apps.

Recently, concatenating multiple keyphrases as a target sequence has been proposed as a new learning paradigm for keyphrase generation. Existing studies concatenate target keyphrases in different orders but no study has examined the effects of ordering on models' behavior. In this paper, we propose several orderings for concatenation and inspect the important factors for training a successful keyphrase generation model. By running comprehensive comparisons, we observe one preferable ordering and summarize a number of empirical findings and challenges, which can shed light on future research on this line of work.

Different texts shall by nature correspond to different number of keyphrases. This desideratum is largely missing from existing neural keyphrase generation models. In this study, we address this problem from both modeling and evaluation perspectives.We first propose a recurrent-generative model that generates multiple keyphrases as delimiter-separated sequences. Generation diversity is further enhanced with two novel techniques by manipulating decoder hidden states. In contrast to previous approaches, our model is capable of generating variable number of diverse keyphrases.We further propose two evaluation metrics tailored towards variable-number generation. We also introduce a new dataset (ST A C KEX) that expand beyond the only existing genre (i.e., academic writing) in keyphrase generation tasks. With both previous and new evaluation metrics, our model outperforms strong baselines on all datasets.

No abstract