Reproduce “DGC-Net: Cross-Modal Dynamic Routing with Global-Guided Compensation in Multimodal Sentiment Analysis” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Linan Zhu, Huaijing Wu, Xiao Han, Yuhao Wang
DGC-Net addresses multimodal sentiment analysis by using other modalities’ utterance-level context to guide local expert selection and adjusting how refined local evidence supplements each modality’s global representation. On CMU-MOSI it achieves mean absolute error of 0.698, versus 0.718 for GLoMo under shared features and partitions. However, gains vary by metric: CMU-MOSI seven-class and negative/positive accuracy and F1 are lower.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.