Research on the Assessment System of English Learners' Intercultural Communication Competence Based on Deep Learning

Abstract

As the core objective of English language teaching, the construction of a scientific assessment system for intercultural communicative competence faces multiple challenges. Traditional assessment methods are generally characterized by single dimensions and strong subjectivity, making it difficult to comprehensively reflect learners' comprehensive performance in real communicative scenarios. In this paper, an intercultural communicative competence assessment system with multidimensional diagnostic ability is constructed by integrating deep learning and traditional feature analysis methods, providing a breakthrough assessment tool for English teaching. The system adopts a multi-task transfer learning architecture, which significantly improves the model's ability to parse cross-cultural contexts by sharing the underlying feature representations of different assessment dimensions. Experiments show that the system's accuracy in determining the appropriateness of rejection strategies reaches 93.4%, which is better than the 67.2% of the traditional questionnaire method. The application of the system in real teaching scenarios verifies its innovative value, and the teaching group that adopts the feedback from this system accelerates the speed of cross-cultural sensitivity improvement of its learners by 2.3 times. Through the fusion mechanism of deep learning and traditional feature extraction, the logical framework of intercultural communicative competence assessment is significantly reconfigured, and the interpretability and generalization ability of the model is improved.