analiza aiqbench
This research proposes a hybrid architecture where Large Language Models act as intent compilers for cloud-edge resource management, decoupling high-level decision-making from low-latency deterministic control loops. The study addresses catastrophic forgetting in sequential tuning by scoring model outputs against execution results and utilizing replay mechanisms to maintain robustness across drifting infrastructure contexts.
Continual Instruction Tuning for LLM-ased Management of Continuum Resources
De la arXiv
Text integralarXiv:2611.10292v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly explored for resource management across the Cloud-Edge continuum. However, embedding LLMs directly inside latency-sensitive control loops places inference latency, execution jitter, and hallucination risks on the critical path. This study argues for a different use of LLMs: a continually instruction-tuned intent compiler that runs once per management intent or declared context change, above a fast, deterministic tactical controller that handles all low-latency runtime actions.
We formulate intent compilation as a continual instruction tuning problem, because management intents arrive sequentially and accumulate while infrastructure and workload contexts drift. Unlike recent LLM-based management frameworks, our design scores the model's outputs by execution, so that catastrophic forgetting can be measured in the executed objective of the underlying controller. On a continuum simulator with a 2B backbone and parameter-efficient adapters, preliminary results show that naive sequential tuning forgets earlier intents in the executed objective, that replay of stored examples nearly prevents this forgetting, and that the tuned compact compiler performs on par with simpler translators while being more robust than a grammar parser to rephrased intents.
Abstract:Large language models (LLMs) are increasingly explored for resource management across the Cloud-Edge continuum. However, embedding LLMs directly inside latency-sensitive control loops places inference latency, execution jitter, and hallucination risks on the critical path. This study argues for a different use of LLMs: a continually instruction-tuned intent compiler that runs once per management intent or declared context change, above a fast, deterministic tactical controller that handles all low-latency runtime actions.
We formulate intent compilation as a continual instruction tuning problem, because management intents arrive sequentially and accumulate while infrastructure and workload contexts drift. Unlike recent LLM-based management frameworks, our design scores the model's outputs by execution, so that catastrophic forgetting can be measured in the executed objective of the underlying controller. On a continuum simulator with a 2B backbone and parameter-efficient adapters, preliminary results show that naive sequential tuning forgets earlier intents in the executed objective, that replay of stored examples nearly prevents this forgetting, and that the tuned compact compiler performs on par with simpler translators while being more robust than a grammar parser to rephrased intents.
Accepted at the Workshop on Research and Innovation in Scalable and Efficient Deep Learning (RISE-DL 2026), co-located with the 16th International Conference on the Internet of Things (IoT 2026), Newcastle upon Tyne, UK, November 2026. The Version of Record will appear in the ACM proceedings of IoT 2026
Distributed, Parallel, and Cluster Computing (cs.DC); Machine Learning (cs.LG)
C.2.4; I.2.6; I.2.7
arXiv:2611.10292 [cs.DC]
(or arXiv:2611.10292v2 [cs.DC] for this version)
https://doi.org/10.1145/3860552.3860597 DOI(s) linking to related resources
Text integral sub licența deschisă a publicației · 252 cuvinte
Subiecte conexe

Customize answers per tenant in Amazon Bedrock Managed Knowledge Base
AWS · 16 nov. 2026
Instruct-FD: Can Your Full-Duplex Speech System Follow Turn-Taking Instructions?
arXiv · 25 iul. 2026
DuplexSpeechBench-IFEval: Evaluating Implicit Instruction Following in Full-Duplex Voice Agents
arXiv · 5 sept. 2026

Amazon Bedrock AgentCore harness adds Strands harness defaults, built-in tools, and lifecycle hooks
AWS · 19 nov. 2026