Linqi Song
Papers
1
Total Citations
4
H-Index
1
About
Linqi Song is a leading researcher in distributed machine learning, multi-robot systems, and communication-efficient optimization. Her work addresses critical bottlenecks in collaborative AI, particularly how to train models across networks of robots or edge devices with limited bandwidth. Her most cited paper, "Adaptive Top-K in SGD for Communication-Efficient Distributed Learning in Multi-Robot Collaboration" (2024, 4 citations), introduces an innovative gradient compression technique that dynamically adjusts the number of gradients transmitted during distributed stochastic gradient descent. This adaptive Top-K sparsification method significantly reduces communication overhead while maintaining model accuracy—a breakthrough for real-world multi-robot coordination. Song’s contributions bridge theoretical optimization with practical deployment, enabling faster, more scalable learning in bandwidth-constrained environments. Her research has direct applications in autonomous swarms, IoT networks, and federated learning systems. With a growing citation footprint, Song is recognized for tackling fundamental trade-offs between communication cost and learning performance. Her work continues to shape how distributed intelligence systems collaborate efficiently, making her a rising voice in the intersection of machine learning, robotics, and networked systems.
Research Focus
Key Achievements
Top Papers
- 1