Fengtao Sun
Papers
1
Total Citations
1
H-Index
1
About
Fengtao Sun is a rising researcher in the field of multimodal machine learning, with a focus on weakly supervised learning and affordance grounding. His work bridges computer vision and natural language processing to enable machines to understand how objects can be used—a key step toward more intuitive human-robot interaction and assistive technologies. Sun’s most notable contribution, "MACR-afford," introduces a novel framework that combines multi-branch attention enhancement with Chain-of-Thought (CoT) multi-stage reasoning, allowing models to infer object affordances from limited or noisy visual and textual data. This approach significantly advances weakly supervised affordance grounding, reducing the need for expensive, fully annotated datasets. While his citation count is still growing, the 2025 publication of MACR-afford signals a promising trajectory in a cutting-edge area of AI. Sun’s work is particularly relevant for researchers in robotics, autonomous systems, and human-computer interaction, as it offers a scalable pathway for machines to learn functional object knowledge. His innovative integration of reasoning chains with attention mechanisms represents a meaningful step toward more explainable and capable AI systems.
Research Focus
Key Achievements
Top Papers
- 1