Baolin Peng

Papers

Total Citations

H-Index

About

Baolin Peng is a leading researcher at the intersection of artificial intelligence, multimodal learning, and social intelligence. His work focuses on building foundation models that bridge digital and physical worlds, with a particular emphasis on developing AI agents capable of understanding and acting across multiple modalities. Peng's most notable contribution is the development of **Magma**, a groundbreaking foundation model for multimodal AI agents that extends traditional vision-language models by equipping them with both verbal intelligence and agentic capabilities for real-world tasks. This work, published in 2025, has already garnered 19 citations, signaling its immediate impact on the field of embodied AI. In parallel, Peng has made significant strides in social relation inference through his work on **SocAoG**, an incremental graph parsing approach that models social networks as And-or Graphs to infer relationships from dialogue. This research, with 12 citations, demonstrates his commitment to building emotionally intelligent systems that can interpret human social dynamics. Peng's dual focus on multimodal agency and social cognition positions him at the forefront of creating AI that can both act intelligently and understand human social contexts, making his work essential reading for researchers in human-AI interaction and autonomous systems.