Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

JAUNT: Joint Alignment of User Intent and Network State for QoE-centric LLM Tool Routing

Enhan Li, Hongyang Du

TL;DR

JAUNT addresses the gap in LLM tool routing by jointly aligning user intent with dynamic network state to maximize QoE. It introduces the TRIP benchmark to model diverse user profiles, emotions, and network variability, and presents a three-module JAUNT framework comprising Semantic Intent Inference, Network Latency Prediction, and Joint QoE-centric Tool Routing. The QoE model combines latency-sensitive degradation via $D(L)=w_1\ln\left(1+\frac{L}{L_{\text{th}}}\right)$ with a success component modulated by $w_2$, and a multiplicative success probability $p_s(L,x;q,u)$ across routing, tool, and network channels. Experimental results show JAUNT achieving higher QoE than baselines under both smooth and random network conditions, validating the importance of jointly modeling user intent and network dynamics for scalable, user-centric LLM orchestration. The work lays groundwork for adaptive, emotion-aware, network-aware routing and suggests future directions toward multi-agent, cross-platform coordination for real-time LLM services.

Abstract

Large Language Models (LLMs) increasingly rely on emerging protocols such as the Model Context Protocol (MCP) to invoke external tools and services. However, current tool routing mechanisms remain fragile because they only consider functional matching between users' queries and tools. In practice, user intent expressed through queries can be vague or underspecified, and the actual Quality of Experience (QoE) also depends on external factors such as link latency and server availability that are not captured by semantics alone. To address this challenge, we propose JAUNT, a framework for Joint Alignment of User intent and Network state in QoE-centric Tool routing. JAUNT introduces a dual-view alignment strategy that interprets user intent while employing LLM agents to construct network profiles, mapping numerical performance indicators into the semantic space to guide routing. We further design a benchmark that integrates diverse user request patterns with heterogeneous network states, enabling systematic evaluation of QoE outcomes. Experimental results show that JAUNT significantly improves QoE compared with several baselines, demonstrating the importance of aligning both intent and network state for scalable LLM service orchestration.

JAUNT: Joint Alignment of User Intent and Network State for QoE-centric LLM Tool Routing

TL;DR

JAUNT addresses the gap in LLM tool routing by jointly aligning user intent with dynamic network state to maximize QoE. It introduces the TRIP benchmark to model diverse user profiles, emotions, and network variability, and presents a three-module JAUNT framework comprising Semantic Intent Inference, Network Latency Prediction, and Joint QoE-centric Tool Routing. The QoE model combines latency-sensitive degradation via with a success component modulated by , and a multiplicative success probability across routing, tool, and network channels. Experimental results show JAUNT achieving higher QoE than baselines under both smooth and random network conditions, validating the importance of jointly modeling user intent and network dynamics for scalable, user-centric LLM orchestration. The work lays groundwork for adaptive, emotion-aware, network-aware routing and suggests future directions toward multi-agent, cross-platform coordination for real-time LLM services.

Abstract

Large Language Models (LLMs) increasingly rely on emerging protocols such as the Model Context Protocol (MCP) to invoke external tools and services. However, current tool routing mechanisms remain fragile because they only consider functional matching between users' queries and tools. In practice, user intent expressed through queries can be vague or underspecified, and the actual Quality of Experience (QoE) also depends on external factors such as link latency and server availability that are not captured by semantics alone. To address this challenge, we propose JAUNT, a framework for Joint Alignment of User intent and Network state in QoE-centric Tool routing. JAUNT introduces a dual-view alignment strategy that interprets user intent while employing LLM agents to construct network profiles, mapping numerical performance indicators into the semantic space to guide routing. We further design a benchmark that integrates diverse user request patterns with heterogeneous network states, enabling systematic evaluation of QoE outcomes. Experimental results show that JAUNT significantly improves QoE compared with several baselines, demonstrating the importance of aligning both intent and network state for scalable LLM service orchestration.
Paper Structure (18 sections, 11 equations, 8 figures, 1 table, 1 algorithm)

This paper contains 18 sections, 11 equations, 8 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. LLM Tool Routing and Evaluation
  4. Quality of Experience
  5. User-Centric QoE Modeling
  6. TRIP Benchmark Construction
  7. User Profile Generation
  8. Clear Query Generation
  9. Semantic Ambiguity Modeling
  10. Emotional Expression Modeling
  11. JAUNT
  12. Semantic Intent Inference Module
  13. Network Latency Prediction Module
  14. Joint QoE-centric Tool Routing Module
  15. Experiment
  16. ...and 3 more sections

Figures (8)

  • Figure 1: Motivation of JAUNT: Intent is obscured by semantic ambiguity, emotional expression, and network trade-offs.
  • Figure 2: Overview of TRIP benchmark construction pipeline.
  • Figure 3: Overall architecture of the JAUNT framework.
  • Figure 4: User profile update mechanism. The profile is dynamically refined by extracting tone and implicit preferences from real-time queries, thereby enhancing the user's QoE.
  • Figure 5: Two network scenarios to evaluate JAUNT's performance under different conditions.
  • ...and 3 more figures