This study examines the double-edged sword effect of chatbot agent anthropomorphism on customer acceptance in retail e-commerce using social response theory. Scenario-based experiments with a between-subject design were conducted to test hypotheses. Study 1 investigated the differences in positive and negative psychological perceptions (i.e. perceived competence, privacy concerns) and acceptance intention among 138 participants for different anthropomorphism levels of chatbots. Study 2 randomly assigned 282 subjects to 2 (anthropomorphism: high vs. low) × 2 (time pressure: high vs. low) experiments. Data collected from users of chatbot services in the retail e-commerce platforms were analysed using methods of reliability and validity analyses, one-sample t-test, independent-sample t-test, one-way ANOVA, and PROCESS bootstrap analyses. The results showed that chatbot anthropomorphism increased customers’ perceptions of chatbot competence and thus increased their intention to accept chatbot services, but also increased customers’ privacy concerns and thus decreased acceptance intention. Under high (vs. low) time pressure, customers perceived highly anthropomorphic chatbots as more competent, but time pressure did not moderate the effect of anthropomorphism on privacy concerns. This study extends research on social response theory and time pressure in the e-commerce management field, and provides guidance for retail e-commerce enterprises to develop chatbot anthropomorphism level differentiation strategies.

Concept prerequisite learning (CPL) plays a key role in developing technologies that assist people to learn a new complex topic or concept.Previous work commonly assumes that all concepts are given at training time and solely focuses on predicting the unseen prerequisite relationships between them.However, many real-world scenarios deal with concepts that are left undiscovered at training time, which is relatively unexplored.This paper studies this problem and proposes a novel alternating knowledge distillation approach to take advantage of both content-and graph-based models for this task.Extensive experiments on three public benchmarks demonstrate up to 10% improvements in terms of F1 score.

Xin Dong, Yaxin Zhu, Zuohui Fu, Dongkuan Xu, Gerard de Melo. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Yaxin Zhu, Yikun Xian, Zuohui Fu, Gerard de Melo, Yongfeng Zhang. Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2021.

There is increasing recognition of the need for human-centered AI that learns from human feedback. However, most current AI systems focus more on the model design, but less on human participation as part of the pipeline. In this work, we propose a Human-in-the-Loop (HitL) graph reasoning paradigm and develop a corresponding dataset named HOOPS for the task of KG-driven conversational recommendation. Specifically, we first construct a KG interpreting diverse user behaviors and identify pertinent attribute entities for each user--item pair. Then we simulate the conversational turns reflecting the human decision making process of choosing suitable items tracing the KG structures transparently. We also provide a benchmark method with reported performance on the dataset to ascertain the feasibility of HitL graph reasoning for recommendation using our developed dataset, and show that it provides novel opportunities for the research community.

In Dunhuang Mogao Grottoes, unique Buddhist murals of ancient China are preserved. Unfortunately, the exquisite murals are suffering from degradation. Experts have been trying to enhance public's awareness of mural protection, but there's no efficacious means to attract interest and popularize knowledge yet. In this paper, we propose RestoreVR, an interactive virtual reality (VR) system engaging users to experience Dunhuang mural restoration in a digital tour in the cave. Based on an online survey with the public and in-depth interviews with five Dunhuang experts, we derive a set of design requirements for generating embodied knowledge and situated experience in VR to bridge the gap between highly specialized experts and general audiences. Accordingly, we design RestoreVR and conduct a between-subjects user study to compare our system with traditional methods. The results suggest that RestoreVR significantly improves user experience and awareness of CH protection over existing methods.

This paper presents LinkBricks, a creative construction kit for intuitively creating and programming interactive robots towards young children. Integrating building blocks, a hierarchical programming framework and a tablet application, this kit is proposed to maintain the low floor and wide walls for children who lack knowledge in conventional programming. The blocks have LEGO-compatible interlock structures and are embedded with various wireless sensors and actuators to create different interactive robots. The programming application is easy-to-use and provides heuristics to involve children in the creative activities. A preliminary evaluation is conducted and indicates that LinkBricks increases young children's engagement with, comfort with, and interest in working with interactive robots. Meanwhile, it has the potential of helping them to learn the concepts of programming and robots.

In cross-lingual text classification, one seeks to exploit labeled data from one language to train a text classification model that can then be applied to a completely different language. Recent multilingual representation models have made it much easier to achieve this. Still, there may still be subtle differences between languages that are neglected when doing so. To address this, we present a semi- supervised adversarial training process that minimizes the maximal loss for label-preserving input perturbations. The resulting model then serves as a teacher to induce labels for unlabeled target lan- guage samples that can be used during further adversarial training, allowing us to gradually adapt our model to the target language. Compared with a number of strong baselines, we observe signifi- cant gains in effectiveness on document and intent classification for a diverse set of languages.