Bowei He's paper scales mobile agents via density and collective intelligence
ACM MobiSys 2026 paper proposes two-pronged approach for edge AI scalability.
Mobile agent systems are emerging as a key paradigm for edge devices and AIoT, but face scalability limits from constrained on-device computation and fragmented intelligence. Bowei He's paper proposes a unified research agenda addressing two complementary dimensions: improving capability density of individual agents through compact foundation model design and compression, and enabling collective intelligence via communication-rich multi-agent collaboration. This vision leverages recent advances in model and infrastructure technologies.
The agenda aims to transform isolated mobile agents into a distributed intelligent system that is both efficient and scalable. By focusing on both individual agent capabilities and collaborative dynamics, the work provides a roadmap for building practical mobile agent systems that can operate effectively in resource-constrained environments while leveraging collective intelligence from networked peers.
- Improves agent capability density via compact foundation models and compression
- Enables collective intelligence through communication-rich multi-agent collaboration
- Accepted at ACM MobiSys 2026; targets scalable edge AI and AIoT ecosystems
Why It Matters
Legacy edge AI systems falter under fragmentation; this agenda shows how to build scalable, collaborative mobile agents.