Xiang Fan
Papers
3
Total Citations
29
H-Index
3
About
Xiang Fan is a leading researcher in multimodal representation learning, a field focused on integrating and interpreting data from diverse sources such as text, images, audio, and sensor signals. His major contributions center on developing novel frameworks to quantify and model the complex interactions between different modalities—a critical challenge for applications in robotics, affective computing, and healthcare. Fan’s most cited work, "MultiBench: Multiscale Benchmarks for Multimodal Representation Learning" (2021, 22 citations), provides standardized evaluation tools that have become foundational for the community. He further advanced the field with the "High-Modality Multimodal Transformer" (2022), which addresses heterogeneity in high-modality settings, and his 2023 paper on an information decomposition framework offers a rigorous mathematical approach to understanding multimodal interactions. Through these achievements, Fan has established himself as a key figure pushing the boundaries of how machines learn from multiple data streams, with his work cited for its technical depth and practical impact.
Research Focus
Key Achievements
Top Papers
- 1MultiBench: Multiscale Benchmarks for Multimodal Representation Learning22 citations · 2021
- 2
- 3