Papers
11
Total Citations
112
H-Index
6
About
Yiwei Lyu is a versatile researcher whose work spans two compelling frontiers: multimodal machine learning and safe autonomous multi-robot systems. In the realm of AI interpretability and representation learning, Lyu has made notable contributions through projects like DIME, which introduced disentangled local explanations for multimodal models to enhance human understanding of AI decision-making (29 citations), and MultiBench, a large-scale benchmarking suite for multimodal representation learning spanning diverse real-world domains (22 citations). His High-Modality Multimodal Transformer further advances the field by addressing heterogeneity across many simultaneous data modalities. Equally impressive is Lyu's work in robotics and autonomous systems, where he has pioneered safety-critical control frameworks for multi-agent coordination. His research on Parametric Control Barrier Functions for autonomous vehicle merging (18 citations) and decentralized risk-aware control with dynamic responsibility allocation (10 citations) addresses fundamental challenges in scalable, provably safe robot behavior. Additional contributions tackling deadlock avoidance, line-of-sight connectivity maintenance under uncertainty, and human-robot interaction demonstrate the breadth and real-world relevance of his portfolio. With over 110 cumulative citations across diverse domains, Lyu's research uniquely bridges intelligent perception and trustworthy autonomous decision-making.
Research Focus
Key Achievements
Top Papers
- 1
- 2MultiBench: Multiscale Benchmarks for Multimodal Representation Learning22 citations · 2021
- 3
- 4
- 5
- 6Design and Control of A Hybrid Sailboat for Enhanced Tacking Maneuverv6 citations · 2018
- 7
- 8Active Probing and Influencing Human Behaviors Via Autonomous Agents4 citations · 2023
- 9
- 10