Phonological and Semantic Embedding Integration in OpenPangu-MLA for Zero-Shot Emotion Recognition

This report synthesises findings from 2 peer-reviewed papers addressing the following research question: How does the integration of phonological and semantic embeddings in OpenPangu-MLA affect its zero-shot emotion recognition performance compared to prosody-only models on the MMSU benchmark. 9 claims were extracted from source literature; 8 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 7.8/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the integration of phonological and semantic embeddings in OpenPangu-MLA affect its zero-shot emotion recognition performance compared to prosody-only models on the MMSU benchmark?

Autonomous literature synthesis. Automated review score: 7.8/10. Full text and citation available at Assignee Research.