Search results

Fine-tuning pre-trained language models (PLMs), e.g. SciBERT, generally require large numbers of annotated data to achieve state-of-the-art performance on a range of NLP tasks in the scientific domain. However, obtaining fine-tuning data for scientific NLP tasks is still challenging and expensive. In this paper, the authors propose the mix prompt tuning (MPT), which is a semi-supervised method aiming to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks.

Specifically, the proposed method provides multi-perspective representations by combining manually designed prompt templates with automatically learned continuous prompt templates to help the given academic function recognition task take full advantage of knowledge in PLMs. Based on these prompt templates and the fine-tuned PLM, a large number of pseudo labels are assigned to the unlabelled examples. Finally, the authors further fine-tune the PLM using the pseudo training set. The authors evaluate the method on three academic function recognition tasks of different granularity including the citation function, the abstract sentence function and the keyword function, with data sets from the computer science domain and the biomedical domain.

Extensive experiments demonstrate the effectiveness of the method and statistically significant improvements against strong baselines. In particular, it achieves an average increase of 5% in Macro-F1 score compared with fine-tuning, and 6% in Macro-F1 score compared with other semi-supervised methods under low-resource settings.

In addition, MPT is a general method that can be easily applied to other low-resource scientific classification tasks.

This study aims to deepen our understanding of how chatbots’ empathy influences humans–AI relationship in frontline service encounters. The authors investigate the underlying mechanisms, including perceived anthropomorphism, perceived intelligence and psychological empowerment, while also considering variations between different stages of the customer journey (before and after purchase).

Data collection was conducted through an online survey distributed among 301 customers who had experience using AI-based service chatbot in frontline service encounters in China. The hypotheses were examined through structural equation modeling and multi-group analysis.

The findings of this study revealed the positive impacts of emotional and cognitive empathy on humans–AI relationship through perceived anthropomorphism, perceived intelligence and psychological empowerment. Furthermore, this study verified the moderating effect of the customer journey stages, such that the impacts of anthropomorphism and intelligence on humans–AI relationship displayed more strength during the pre- and post-purchase phases, respectively.

This research offers practical implications for companies: recognize and enhance empathy dimensions in AI-based service chatbot to empower human–AI relationships; boost customer empowerment in human–AI interactions; and tailor anthropomorphic features in the pre-purchase stage and improve problem-solving capability in the post-purchase stage to enrich user experiences.

This study extends relationship marketing theory and human–AI interaction frameworks by investigating the underlying mechanisms of the effect of two-dimensional empathy on human–AI relationship. This study also enriches service design theories by revealing the moderating effect of customer journey stages.