Purpose This study serves two purposes: (1) to evaluate the effects of organizational ambidexterity by examining how balanced and combined sales–service configurations chatbots differ in their abilities enhance customer experience patronage (2) apply information boundary theory assess contingent role that chatbot can play adapting customers' personalization–privacy paradox. Design/methodology/approach An online survey artificial intelligence users was conducted, a mixed-methods research...

Manifold structure is important for a data set, and many subspace learning methods tend to preserve this in the process. In paper, we simultaneously consider distances angles between image vectors measure similarities, hope of more sufficiently capturing manifold structure. order highlight distinctions among different data, enhance complementary information compared with distance, propose new type angle measurement shifted space that centered at mean. Both distance are fused using parallel...

Uncovering rationales behind predictions of graph neural networks (GNNs) has received increasing attention over the years. Existing literature mainly focus on selecting a subgraph, through combinatorial optimization, to provide faithful explanations. However, exponential size candidate subgraphs limits applicability state-of-the-art methods large-scale GNNs. We enhance this different approach: by proposing generative structure -- GFlowNets-based GNN Explainer (GFlowExplainer), we turn...

Large Language Models (LLMs) have proven powerful, but the risk of privacy leakage remains a significant concern. Traditional privacy-preserving methods, such as Differential Privacy and Homomorphic Encryption, are inadequate for black-box API-only settings, demanding either model transparency or heavy computational resources. We propose Prompt2Forget (P2F), first framework designed to tackle LLM local challenge by teaching forget. The method involves decomposing full questions into smaller...

Shield construction technology has been widely used in subway construction, and reasonable prediction of shield tunneling parameters is great practical significance for improving safety reducing operational difficulties.Taking a section the Hangzhou Airport Express Underpass Tunnel as engineering background, this study uses friction angle, cohesion force, compression modulus heaviness soil layer within diameter tunnel well burial depth cover, preset blade speed propulsion machine inputs,...

When extracting discriminative features from multimodal data, current methods rarely concern themselves with the data distribution. In this paper, we present an assumption that is consistent viewpoint of discrimination, is, a person’s overall biometric should be regarded as one class in input space, and his different can form Gaussians distributions, i.e., subclasses. Hence, propose novel feature extraction recognition approach based on subclass discriminant analysis (SDA). Specifically,...

With the rapid development of information and Internet technology, online education has become an increasingly popular way education. Online allows learners to learn any content at place time. However, in process learning, learners' learning mood state is usually not paid attention to. Due a long time face monotonous non-communication computer screen, are prone physical or psychological fatigue, resulting decreased efficiency. In view this phenomenon, taking into account characteristics we...

Many score-based active learning methods have been successfully applied to graph-structured data, aiming reduce the number of labels and achieve better performance graph neural networks based on predefined score functions. However, these algorithms struggle learn policy distributions that are proportional rewards limited exploration capabilities. In this paper, we innovatively formulate problem as a generative process, named GFlowGNN, which generates various samples through sequential...

At present, domestic vocabulary teaching, including academic English mostly adopts traditional teaching methods.The current situation is that students with limited have low competence.It teachers who the duty and responsibility to help expand their vocabulary.Guided by Production-oriented approach (POA), this article manages construct design of explores effective ways enhance students' proficiency.This finds under guidance POA theory, can more effectively stimulate internal motivation,...

In this article, a model‐free parallel reinforcement learning method is proposed to solve the suboptimal control problem for Markov jump singularly perturbed systems. First, since fast and slow dynamics coexist in systems, it may lead ill‐conditioned numerical problems during controller design process. Therefore, original system can be decomposed into independent subsystems at different time‐scales by employing reduced order method. addition, two algorithms are designed obtain optimal...

Detecting multimodal misinformation, especially in the form of image-text pairs, is crucial. Obtaining large-scale, high-quality real-world fact-checking datasets for training detectors costly, leading researchers to use synthetic generated by AI technologies. However, generalizability trained on data scenarios remains unclear due distribution gap. To address this, we propose learning from detecting misinformation through two model-agnostic selection methods that match and distributions....

With the advancement of economic globalization, intercultural communication has gradually become a practical need.From perspective communication, this article takes specific examples to analyze embodiment pragmatic failures in daily life, including pragma-linguistic failure and socio-pragmatic failure.After summarizing manifold reasons attribute more too cultural differences.The significance is help students face problems that appear frequently them effectively avoid failures, thus greatly...