Research and Design of a Chinese-French Consecutive Interpreting Assessment System Based on Artificial Intelligence Technology

In today's era, artificial intelligence technology is developing rapidly. Under such a background, translation quality evaluation system relying on deep learning has gradually become a hot research topic in the field of translation research. Traditional manual evaluation methods have their own shortcomings, which are relatively subjective and low evaluation efficiency. The existing automatic assessment indicators such as BLEU and METEOR, which have deficiencies in cultural adaptability and logical coherence, cannot meet the complex needs of Sino-French alternating interpretation. This evaluation system combines multi-modal analysis with cross-language and cultural feature database. Based on the improvement of Transformer model architecture, a hierarchical evaluation mechanism is designed, which significantly improves the objectivity of evaluation and makes the evaluation more comprehensive. The experimental results show that this system has more advantages than the traditional method in the three core indicators of accuracy, fluency and cultural adaptability. The Pearson correlation coefficient between the score of this system and the human score reaches 0.87, which proves that it is effective in cross-language and cultural adaptation scenarios. This study provides an extensible technical framework for the assessment of translation quality, and provides innovative solutions to the challenges of linguistic structural differences and cultural metaphor translation.