Sijie Cheng
Papers
2
Total Citations
21
H-Index
2
About
Sijie Cheng is a researcher at the forefront of artificial intelligence, with key contributions spanning vision-language models and neuromorphic computing. Their work is distinguished by a focus on advancing AI’s cognitive and structural capabilities. Cheng’s highly cited paper, “EgoThink: Evaluating First-Person Perspective Thinking Capability of Vision-Language Models” (2024, 18 citations), introduces a novel evaluation framework that shifts the paradigm from traditional third-person assessments to first-person reasoning in VLMs. This work addresses a critical gap in understanding how AI perceives and interacts with the world from an egocentric viewpoint, offering new benchmarks for embodied AI research. In parallel, Cheng explores the frontiers of brain-inspired computing with “Evolving Connectivity for Recurrent Spiking Neural Networks” (2023, 3 citations), proposing methods to optimize RSNN architectures for complex temporal dynamics. This research holds promise for more efficient, biologically plausible AI systems. By bridging high-level cognitive evaluation and low-level neural architecture design, Cheng’s work demonstrates a rare ability to tackle AI challenges from multiple scales, making significant strides toward more capable and human-like artificial intelligence.
Research Focus
Key Achievements
Top Papers
- 1
- 2Evolving Connectivity for Recurrent Spiking Neural Networks3 citations · 2023