Optimizing NetGPT via Routing-Based Synergy and Reinforcement Learning
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912734037147648 |
|---|---|
| author | Chen, Yuxuan Li, Rongpeng Chen, Xianfu Wu, Celimuge Peng, Chenghui Zhao, Zhifeng Zhang, Honggang |
| author_facet | Chen, Yuxuan Li, Rongpeng Chen, Xianfu Wu, Celimuge Peng, Chenghui Zhao, Zhifeng Zhang, Honggang |
| contents | Large language model (LLM) agents at the network edge offer low-latency execution for routine queries. In contrast, complex requests often require the superior capability of cloud models, incurring higher latency and cost. To navigate this quality-cost trade-off under dynamic network conditions, we propose a cloud-edge synergy for NetGPT that integrates network-aware routing with on-edge self-improvement. Specifically, our framework routes structured tool-calling requests to cloud or edge agents via a novel scoring policy. We prove that, under mild regularity assumptions, the optimal routing rule admits a unique fallback threshold with monotone dependence on bandwidth and round-trip time (RTT). Concurrently, based on the dataset collected from requests routed to the cloud and corresponding responses, we instantiate a schema-preserving reinforcement learning (RL) to improve the capability of the edge agent. We analyze a supervised finetuning (SFT)-anchored composite objective that combines a reverse-KL trust-region step with a forward-KL realignment toward the SFT prior, explaining stability and constraining policy drift. Both the network-aware routing policy and the edge agent are updated coherently. Experiments across controlled network states and pricing schedules demonstrate smooth quality-cost frontiers, consistent gains of dynamic fallback thresholds over fixed policies, and sustained reductions in offloading while maintaining task success and schema-correct outputs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_22217 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Optimizing NetGPT via Routing-Based Synergy and Reinforcement Learning Chen, Yuxuan Li, Rongpeng Chen, Xianfu Wu, Celimuge Peng, Chenghui Zhao, Zhifeng Zhang, Honggang Networking and Internet Architecture Large language model (LLM) agents at the network edge offer low-latency execution for routine queries. In contrast, complex requests often require the superior capability of cloud models, incurring higher latency and cost. To navigate this quality-cost trade-off under dynamic network conditions, we propose a cloud-edge synergy for NetGPT that integrates network-aware routing with on-edge self-improvement. Specifically, our framework routes structured tool-calling requests to cloud or edge agents via a novel scoring policy. We prove that, under mild regularity assumptions, the optimal routing rule admits a unique fallback threshold with monotone dependence on bandwidth and round-trip time (RTT). Concurrently, based on the dataset collected from requests routed to the cloud and corresponding responses, we instantiate a schema-preserving reinforcement learning (RL) to improve the capability of the edge agent. We analyze a supervised finetuning (SFT)-anchored composite objective that combines a reverse-KL trust-region step with a forward-KL realignment toward the SFT prior, explaining stability and constraining policy drift. Both the network-aware routing policy and the edge agent are updated coherently. Experiments across controlled network states and pricing schedules demonstrate smooth quality-cost frontiers, consistent gains of dynamic fallback thresholds over fixed policies, and sustained reductions in offloading while maintaining task success and schema-correct outputs. |
| title | Optimizing NetGPT via Routing-Based Synergy and Reinforcement Learning |
| topic | Networking and Internet Architecture |
| url | https://arxiv.org/abs/2511.22217 |