Max Robotics — Global Service Robot Directory + US Certification

About

Jiashi Feng is a computer vision and machine learning researcher whose work spans future prediction, depth estimation, human motion forecasting, and embodied AI. His research is particularly distinguished by its focus on enabling intelligent systems — such as autonomous vehicles and robots — to anticipate and interpret dynamic environments with greater accuracy and efficiency. Feng's most influential contribution lies in predictive scene understanding, notably his 2017 work on jointly forecasting scene parsing and motion dynamics, which garnered 49 citations and addressed a critical challenge in autonomous navigation. He has further advanced generative modeling for future frame synthesis through difference-guided GANs, and explored egocentric spatial memory architectures that equip agents with human-like environmental awareness. His 2021 research on velocity-to-velocity human motion forecasting extended these themes into structured prediction of human behavior. More recently, Feng has pushed the frontier of foundation models, introducing prompt-based paradigms for metric depth estimation and developing DeeR-VLA, a dynamic inference framework that makes multimodal large language models more computationally efficient for real-world robot execution. Together, these contributions reflect a sustained commitment to bridging perception, prediction, and actionable intelligence in embodied AI systems.